Half a million species of plants could be existing in the world. Classification of plants based on leaf features is a critical job as feature extraction from binary images of leaves may result in duplicate identification. However, leaves are an effective means of differentiating plant species because of their unique characteristics like area, diameter, perimeter, circularity, aspect ratio, solidity, eccentricity, and narrow factor. This paper presents the extraction of plant leaf gas alongside other features from the camera (...) images or a dataset of images by applying a convolutional neural network. The extraction of leaf gas enables identification of the actual level of chlorophyll and nitrogen which may help to interpret future predictions. Our contribution includes the study of texture and geometric features, analyzing ratio of Ch and N in both healthy and dead leaves, and the study of color-based methods via CNN. Several steps are included to obtain the results: image preprocessing, testing, training, enhancement, segmentation, feature extraction, and aggregation of results. A vital contrast of the results can be seen by considering the kind of image, whether a healthy or dead leaf. (shrink)

This paper studies the feature extraction and middle-level expression of Convolutional Neural Network convolutional layer glass broken and cracked at the scene of road traffic accident. The image pyramid is constructed and used as the input of the CNN model, and the convolutional layer road traffic accident scene glass breakage and crack characteristics at each scale in the pyramid are extracted separately, and then the depth descriptors at different image scales are extracted. In order to improve the discriminative (...) power of the depth descriptor, Hellinger kernel and Principal Component Analysis are used to perform nonlinear and linear transformations. Two aggregation strategies based on depth descriptors are proposed to form a global image representation. The classification experiment of the data set shows that Hellinger kernel, PCA transformation, and two aggregation strategies are all conducive to improving the classification accuracy. In addition, the convolutional layer road traffic accident scene glass breaking and crack feature coding can obtain better classification performance than the fully connected layer feature. We conducted dynamic impact tests on plate glass and Polyvinyl Butyral- laminated glass under different boundary conditions and studied the crack propagation and failure process of the glass under different impact speeds. The results show that there are radial cracks and circular cracks on the glass specimens under the impact load; the glass specimens show partial damage under high-speed impact and the characteristics of glass breaking and cracks at the scene of road traffic accidents; the four-frame plate glass is supported by sharp dagger-like fragments. This paper compares the energy absorption of glass specimens under different boundary conditions. The results show that the energy absorption effect of the four-point supporting glass specimen is generally stronger than that of the four-frame supporting glass. (shrink)

In order to improve the integrity of the social network behavior feature extraction results for sports college students, this study proposes to be based on the clustering algorithm. This study analyzes the social network information dissemination mechanism in the field of college students’ sports, obtains the real-time social behavior data in the network environment combined with the analysis results, and processes the obtained social network behavior data from two aspects of data cleaning and de-duplication. Using clustering algorithm to (...) determine the type of social network user behavior, setting the characteristics of social network behavior attributes, and finally through quantitative and standardized processing, get the results of college students’ sports field social network behavior characteristics extraction. The experimental results showed that the completeness of the method feature extraction results improved to 9.93%, and the average extraction time cost was 0.344 s, with high result integrity and obvious advantages in the extraction speed. (shrink)

The idea of the orderly output constraint is compared with recent findings about the representation of vowels in the auditory cortex of an animal model for human speech sound processing (Ohl & Scheich 1997). The comparison allows a critical consideration of the idea of neuronal “feature extractors,” which is of relevance to the noninvariance problem in speech perception.

A computer-based age estimation is a technique that predicts an individual's age based on visual traits derived by analyzing a 2D picture of the individual's face. Age estimation is critical for access control, e-government, and effective human–computer interaction. The other-race effect has the potential to cause techniques designed for white faces to underperform when used in a region with black faces. The outcome is the consequence of intermittent training with faces of the same race and the encoding structure of the (...) trained face images, which is based on the feature extraction technique used. This study contributes to a constructive comparison of three feature-extraction techniques, namely, local binary pattern, Gabor Wavelet, and wavelet transformation, used in the development of a genetic algorithm -artificial neural network -based age estimation system. The feature extraction techniques used are proven to produce a wealth of shape and textural information. The GA-ANN constitutes the age classifier module. The correct classification rate was chosen as the performance metrics in this study. The results demonstrated that the LBP is a more robust representation of the black face than the GW and Wavelet transformations, as evidenced by its accuracy rate of 91.76 compared to 89.41 and 84.71 achieved with the GW and Wavelet transformation age estimation systems, respectively. (shrink)

Neurons repeatedly exposed to similar perceptual experiences fire together and wire together to form ‘meaning kernels’ of concepts. Pulvermueller argues that abstract concepts may be devoid of meaning kernels, because the perceptual experiences that construct abstract concepts are subject to great variation and share few common features. Abstract concept are therefore grounded in the brain through features that belong to ‘meaning halos’, rather than to ‘meaning kernels’.

Ongoing electroencephalography signals are recorded as a mixture of stimulus-elicited EEG, spontaneous EEG and noises, which poses a huge challenge to current data analyzing techniques, especially when different groups of participants are expected to have common or highly correlated brain activities and some individual dynamics. In this study, we proposed a data-driven shared and unshared feature extraction framework based on nonnegative and coupled tensor factorization, which aims to conduct group-level analysis for the EEG signals from major depression disorder (...) patients and healthy controls when freely listening to music. Constrained tensor factorization not only preserves the multilinear structure of the data, but also considers the common and individual components between the data. The proposed framework, combined with music information retrieval, correlation analysis, and hierarchical clustering, facilitated the simultaneous extraction of shared and unshared spatio-temporal-spectral feature patterns between/in MDD and HC groups. Finally, we obtained two shared feature patterns between MDD and HC groups, and obtained totally three individual feature patterns from HC and MDD groups. The results showed that the MDD and HC groups triggered similar brain dynamics when listening to music, but at the same time, MDD patients also brought some changes in brain oscillatory network characteristics along with music perception. These changes may provide some basis for the clinical diagnosis and the treatment of MDD patients. (shrink)

This article suggests an enhancement of the Masek circle model approach usually used to find a trade-off between modeling complexity, algorithm accuracy, and computational time, mainly for embedded systems where the real-time aspect is a high challenge. Moreover, most commercialized systems today frame iris regions by circles. This work led to several novelties: first, in the segmentation process, the corneal reflection removal method based on morphological reconstruction and pixel connectivity was implemented. Second, the picture size reduction was applied according to (...) nearest-neighbor interpolation. Third, the image gradient of the convolved-reduced picture was then generated using four proposed matrices. Fourth, and to reduce the complexity of the traditional method for the detection of the top and lower eyelids, a new method based on the Radon transform and the least squares fitting method was applied. Fifth, eyelashes were detected via the diagonal gradient and thresholding method. Monogenic signal was used in the feature extraction process. Finally, two distance measures were selected as a metric for recognition. Our experimental results using CASIA iris database V3.0 reveal that the proposed method provides a high performance in terms of speed and accuracy. Using dissimilarity modified Hamming distance, the accuracy of iris recognition was improved, with a false acceptance rate equal to 3% and a speed at least eight times as compared with the state of the art. (shrink)

In speaker recognition systems, feature extraction is a challenging task under environment noise conditions. To improve the robustness of the feature, we proposed a multiscale chaotic feature for speaker recognition. We use a multiresolution analysis technique to capture more finer information on different speakers in the frequency domain. Then, we extracted the speech chaotic characteristics based on the nonlinear dynamic model, which helps to improve the discrimination of features. Finally, we use a GMM-UBM model to develop (...) a speaker recognition system. Our experimental results verified its good performance. Under clean speech and noise speech conditions, the ERR value of our method is reduced by 13.94% and 26.5% compared with the state-of-the-art method, respectively. (shrink)

In face recognition systems, highly robust facial feature representation and good classification algorithm performance can affect the effect of face recognition under unrestricted conditions. To explore the anti-interference performance of convolutional neural network reconstructed by deep learning framework in face image feature extraction and recognition, in the paper, first, the inception structure in the GoogleNet network and the residual error in the ResNet network structure are combined to construct a new deep reconstruction network algorithm, with the random (...) gradient descent and triplet loss functions as the model optimizer and classifier, respectively, and it is applied to the face recognition in Labeled Faces in the Wild face database. Then, the portrait pyramid segmentation and local feature point segmentation are applied to extract the features of face images, and the matching of face feature points is achieved using Euclidean distance and joint Bayesian method. Finally, Matlab software is used to simulate the algorithm proposed in this paper and compare it with other algorithms. The results show that the proposed algorithm has the best face recognition effect when the learning rate is 0.0004, the attenuation coefficient is 0.0001, the training method is SGD, and dropout is 0.1, and the algorithm proposed in this paper has the largest mean average precision compared to other CNN algorithms. The correct rate of face feature matching of the algorithm proposed in this paper is 84.72%, which is higher than LetNet-5, VGG-16, and VGG-19 algorithms, the correct rates of which are 6.94%, 2.5%, and 1.11%, respectively, but lower than GoogleNet, AlexNet, and ResNet algorithms. At the same time, the algorithm proposed in this paper has a faster matching time and a higher correct matching rate than the joint Bayesian method, indicating that the deep reconstruction network algorithm proposed in this paper can be used in face image recognition, FE, and matching, and it has strong anti-interference. (shrink)

The accurate positioning of target is an important link in robot technology. Based on machine learning algorithm, this study firstly analyzed the location positioning principle of binocular vision of robot, then extracted features of the target using speeded-up robust features (SURF) method, positioned the location using Back Propagation Neural Networks (BPNN) method, and tested the method through experiments. The experimental results showed that the feature extraction of SURF method was fast, about 0.2 s, and was less affected by (...) noise. It was found from the positioning results that the output position of the BPNN method was basically consistent with the actual position, and errors in X, Y and Z directions were very small, which could meet the positioning needs of the robot. The experimental results verify the effectiveness of machine learning method and provide some theoretical support for its further promotion and application in practice. (shrink)

Millions of memes are created and shared every day on social media platforms. Memes are a great tool to spread humour. However, some people use it to target an individual or a group generating offensive content in a polite and sarcastic way. Lack of moderation of such memes spreads hatred and can lead to depression like psychological conditions. Many successful studies related to analysis of language such as sentiment analysis and analysis of images such as image classification have been performed. (...) However, most of these studies rely only upon either one of these components. As classifying meme is one problem which cannot be solved by relying upon only any one of these aspects, the present work identifies, addresses, and ensembles both the aspects for analyzing such data. In this research, we propose a solution to the problems in which the classification depends on more than one model. This paper proposes two different approaches to solve the problem of identifying hate memes. The first approach uses sentiment analysis based on image captioning and text written on the meme. The second approach is to combine features from different modalities. These approaches utilize a combination of glove, encoder-decoder, and OCR with Adamax optimizer deep learning algorithms. Facebook Challenge Hateful Meme Dataset is utilized which contains approximately 8500 meme images. Both the approaches are implemented on the live challenge competition by Facebook and predicted quite acceptable results. Both approaches are tested on the validation dataset, and results are found to be promising for both models. (shrink)

High-dimensional data analysis has become the most challenging task nowadays. Dimensionality reduction plays an important role here. It focuses on data features, which have proved their impact on accuracy, execution time, and space requirement. In this study, a dimensionality reduction method is proposed based on the convolution of input features. The experiments are carried out on minimal preprocessed nine benchmark datasets. Results show that the proposed method gives an average 38% feature reduction in the original dimensions. The algorithm accuracy (...) is tested using the decision tree, support vector machine, and K-nearest neighbor classifiers and evaluated with the existing principal component analysis algorithm. The average increase in accuracy is 8.06 for DT, 5.80 for SVM, and 18.80 for the KNN algorithm. The most significant characteristic feature of the proposed model is that it reduces attributes, leading to less computation time without loss in classifier accuracy. (shrink)

Handwritten text recognition is considered as the most challenging task for the research community due to slight change in different characters’ shape in handwritten documents. The unavailability of a standard dataset makes it vaguer in nature for the researchers to work on. To address these problems, this paper presents an optical character recognition system for the recognition of offline Pashto characters. The problem of the unavailability of a standard handwritten Pashto characters database is addressed by developing a medium-sized database of (...) offline Pashto characters. This database consists of 11352 character images. Enriched feature extraction techniques of histogram of oriented gradients and zoning-based density features are used for feature extraction of carved Pashto characters. K-nearest neighbors is considered as a classification tool for the proposed algorithm based on the proposed feature sets. A resultant accuracy of 80.34% is calculated for the histogram of oriented gradients, while for zoning-based density features, 76.42% is achieved using 10-fold cross validation. (shrink)

A session-based recommendation system is designed to predict the user’s next click behavior based on an ongoing session. Existing session-based recommendation systems usually model a session into a sequence and extract sequence features through recurrent neural network. Although the performance is greatly improved, these procedures ignore the relationships between items that contain rich information. In order to obtain rich items embeddings, we propose a novel Recommendation Model based on Multi-channel Convolutional Neural Network for session-based recommendation, RMMCNN for brevity. Specifically, we (...) capture items' internal features from three dimensions through multi-channel convolutional neural network firstly. Next, we merge the internal features with external features obtained by a GRU unit. Then, both internal features and external features are merged by an attention mechanism together as the input of the transformation function. Finally, the probability distribution is taken as the output after the softmax function. Experiments on various datasets show that our method's precision and recommendation performance are better than those of other state-of-the-art approaches. (shrink)

Color harmony is the focus of many researchers in the field of art and design, and its research results have been widely used in artistic creation and design activities. With the development of signal processing and artificial intelligence technology, new ideas and methods are provided for color harmony theory and color harmony calculation. In this article, psychological experimental methods and information technology are combined to design and quantify the 16-dimensional physical features of multiple colors, including multi-color statistical features and multi-color (...) contrast features. Eighty-four subjects are invited to give a 5-level score on the degree of color harmony for 164 multi-color materials selected from the screenshots of film and television scenes. Based on the multi-color physical features and the subjective evaluation experiment, the correlation analysis is firstly carried out, which shows that the overall lightness, difference of the color tones, number of multiple colors, lightness contrast, color tone contrast, and cool/warm contrast are significantly correlated with color harmony. On the other hand, the regression prediction model and classification prediction model of color harmony are constructed based on machine learning algorithms. In terms of regression prediction model, the prediction accuracy of linear models is higher than that of nonlinear models, with 63.9% as the highest, indicating that the multi-color physical features can explain color harmony well. In terms of classification prediction model, the Random Forest has the best prediction performance, with an accuracy of 80.2%. (shrink)

How to strengthen emergency management and improve the ability to prevent and respond to emergencies is an important part of building a harmonious socialist society. This paper proposes a domain emotion dictionary construction method for network public opinion analysis of public emergencies. Using the advantages of corpus and semantic knowledge base, this paper extracts the seed words based on the large-scale network public opinion corpus and combined with the existing emotion dictionary, trains the word vector through the word2vec model in (...) deep learning, expands the emotion words, and obtains the candidate emotion words according to the semantic similarity calculation, So as to generate a domain emotion dictionary. The accuracy rate of emotion discrimination by the emotion dictionary constructed in this paper is 0.86, the recall rate is 0.92. Through the verification of accuracy and recall rate, the construction method proposed in this paper has good accuracy and reliability. Because of the great differences in different experiences and situations of different groups, there will be great differences in views and perspectives on the same event. The key to prevent the public from blindly following the crowd should be to reach groups close to emotional distance, and targeted prevention and control of public opinion can be conducted according to different characteristics of different groups. (shrink)

How speech sounds are represented in the brain is not fully understood. The mismatch negativity has proven to be a powerful tool in this regard. The MMN event-related potential is elicited by a deviant stimulus embedded within a series of repeating standard stimuli. Listeners construct auditory memory representations of these standards despite acoustic variability. In most designs that test speech sounds, however, this variation is typically intra-category: All standards belong to the same phonetic category. In the current paper, inter-category variation (...) is presented in the standards. These standards vary in manner of articulation but share a common phonetic feature. In the standard retroflex experimental block, Mandarin Chinese speaking participants are presented with a series of “standard” consonants that share the feature [retroflex], interrupted by infrequent non-retroflex deviants. In the non-retroflex standard experimental block, non-retroflex standards are interrupted by infrequent retroflex deviants. The within-block MMN was calculated, as was the identity MMN to account for intrinsic differences in responses to the stimuli. We only observed a within-block MMN to the non-retroflex deviant embedded in the standard retroflex block. This suggests that listeners extract [retroflex] despite significant inter-category variation. In the non-retroflex standard block, because there is little on which to base a coherent auditory memory representation, no within-block MMN was observed. The iMMN to the retroflex was observed in a late time-window at centro-parieto-occipital electrode sites instead of fronto-central electrodes, where the MMN is typically observed, potentially reflecting the increased difficulty posed by the added variation in the standards. In short, participants can construct auditory memory representations despite significant acoustic and inter-category phonological variation so long as a shared phonetic feature binds them together. (shrink)

With the rapid emergence of the technology of deep learning, it was successfully used in different fields such as the aquatic product. New opportunities in addition to challenges can be created according to this change for helping data processing in the smart fish farm. This study focuses on deep learning applications and how to support different activities in aquatic like identification of the fish, species classification, feeding decision, behavior analysis, estimation size, and prediction of water quality. Power and performance of (...) computing with the analyzed given data are applied in the proposed DL method within fish farming. Results of the proposed method show the significance of contributions in deep learning and how automatic features are extracted. Still, there is a big challenge of using deep learning in an era of artificial intelligence. Training of the proposed method used a large number of labeled images got from the Fish4Knowledge dataset. The proposed method based on suitable feature extracted from the fish achieved good results in terms of recognition rate and accuracy. (shrink)

This article aims to explore representation of the content knowledge of historical Malay manuscripts by extracting the event features using an event ontology framework. The manuscript used during the testing is Sulalatus Salatin by Abdul Ahmad Samad and it was published at University of Malaya Digital Library database. In aligning to a domain-specific ontology, the Simple Event Model model is adopted and an event-based ontology for historical Malay manuscripts is designed. Information extraction approach is done manually to extract events (...) from the manuscript and mapped into Protégé editor. Competency questions were constructed and submitted to the Protégé editor using SPARQL to check the ontology capability of providing answers as well as to examine its correctness. Event-based ontology model assists in discovering and representing the content knowledge of historical Malay manuscripts and supports organisation of knowledge. All the main concepts are extracted from selected Malay manuscript and 17 concepts used to develop the event-based ontology model. The knowledge was verified by three domain experts in Malay manuscript. In the findings, the interrater reliability for Event and Actor instances is 84%, which means 16% of instances and its type are incorrect and need amendment. For Place, interrater reliability is 95% and 99% for Role. Meanwhile, the experts achieved 100% agreement for Time. In addition, the experts agreed that the concepts, properties and instances for Malay Manuscript Ontology and complied with the criteria of consistency, completeness, conciseness, expandability and ease of use. The development of the event-based model of an ontology-based system with a high level of semantic granularity reflects the various cultural riches and intellectual aspect stored in Malay manuscripts. This will enable systematic research of the knowledge embedded in the manuscripts and make it widely and easily accessible by everyone. (shrink)

Keyword extraction refers to the process of selecting most significant, relevant, and descriptive terms as keywords, which are present inside a single document. Keyword extraction has major applications in the information retrieval domain, such as analysis, summarization, indexing, and search, of documents. In this paper, we present a novel supervised technique for extraction of keywords from medium-sized documents, namely Corpus-based Contextual Semantic Smoothing. CCSS extends the concept of Contextual Semantic Smoothing, which considers term usage patterns in similar (...) texts to improve term relevance information. We introduce four more features beyond CSS as our novel contributions in this work. We systematically compare the performance of CCSS with other techniques, when implemented over INSPEC dataset, where CCSS outperforms all state-of-the-art keyphrase extraction techniques presented in the literature. (shrink)

We describe research carried out as part of a text summarisation project for the legal domain for which we use a new XML corpus of judgments of the UK House of Lords. These judgments represent a particularly important part of public discourse due to the role that precedents play in English law. We present experimental results using a range of features and machine learning techniques for the task of predicting the rhetorical status of sentences and for the task of selecting (...) the most summary-worthy sentences from a document. Results for these components are encouraging as they achieve state-of-the-art accuracy using robust, automatically generated cue phrase information. Sample output from the system illustrates the potential of summarisation technology for legal information management systems and highlights the utility of our rhetorical annotation scheme as a model of legal discourse, which provides a clear means for structuring summaries and tailoring them to different types of users. (shrink)

Convolutional neural networks (CNNs) are increasingly widely used in psychology and neuroscience to predict how human minds and brains respond to visual images. Typically, CNNs represent these images using thousands of features that are learned through extensive training on image datasets. This raises a question: How many of these features are really needed to model human behavior? Here, we attempt to estimate the number of dimensions in CNN representations that are required to capture human psychological representations in two ways: (1) (...) directly, using human similarity judgments and (2) indirectly, in the context of categorization. In both cases, we find that low-dimensional projections of CNN representations are sufficient to predict human behavior. We show that these low-dimensional representations can be easily interpreted, providing further insight into how people represent visual information. A series of control studies indicate that these findings are not due to the size of the dataset we used and may be due to a high level of redundancy in the features appearing in CNN representations. (shrink)

Traditional methods for deriving property-based representations of concepts from text have focused on either extracting only a subset of possible relation types, such as hyponymy/hypernymy (e.g., car is-a vehicle) or meronymy/metonymy (e.g., car has wheels), or unspecified relations (e.g., car—petrol). We propose a system for the challenging task of automatic, large-scale acquisition of unconstrained, human-like property norms from large text corpora, and discuss the theoretical implications of such a system. We employ syntactic, semantic, and encyclopedic information to guide our (...) class='Hi'>extraction, yielding concept-relation-feature triples (e.g., car be fast, car require petrol, car cause pollution), which approximate property-based conceptual representations. Our novel method extracts candidate triples from parsed corpora (Wikipedia and the British National Corpus) using syntactically and grammatically motivated rules, then reweights triples with a linear combination of their frequency and four statistical metrics. We assess our system output in three ways: lexical comparison with norms derived from human-generated property norm data, direct evaluation by four human judges, and a semantic distance comparison with both WordNet similarity data and human-judged concept similarity ratings. Our system offers a viable and performant method of plausible triple extraction: Our lexical comparison shows comparable performance to the current state-of-the-art, while subsequent evaluations exhibit the human-like character of our generated properties. (shrink)

In this paper, a novel approach for word extraction and character segmentation from the handwritten Bangla document images is reported. At first, a modified Run Length Smoothing Algorithm, called Spiral Run Length Smearing Algorithm, is applied for the extraction of words from the text lines of unconstrained handwritten Bangla document images. This technique has helped to overcome some of the drawbacks of standard horizontal and vertical RLSA techniques. SRLSA technique has been applied on the Bangla handwritten document image (...) database CMATERdb1.1.1 and the success rate of the word extraction is found to be 86.01%. In the second part of the work, we have presented a useful solution to the problem on how best word images of handwritten Bangla script can be segmented into constituent characters. Moreover, the technique can segment the words having discontinuity in Matra, a prominent feature of Bangla script. It also optimizes the trade-off between under/over segmentation as Matra region and segmentation points are estimated more precisely. As a result, better word segmentation accuracy is achieved with minimal data loss. Here, a success rate of 92.48% is observed on a dataset of 750 handwritten Bangla words which is 3.35% higher than that of our earlier techniques. (shrink)

Extractive reserves, usually associated with the survival of rubber tappers in the Brazilian tropics, have close parallels elsewhere, including temperate zones. This research isolates the distinctive features of recent Amazonian reserves, illustrates parallel features in a fifty year-old management experiment in the United States, and explores the advantages extractive reserves offer land reformers interested not only in social equity and efficiency but in biological conservation. Extractive reserves stand apart from traditional land reforms in their innovative use of common property, a (...) tenure mode well adapted to sustainable management of marginal lands. (shrink)

Decisions of regulatory government bodies and courts affect many aspects of citizens’ lives. These organizations and courts are expected to provide timely and coherent decisions, although they struggle to keep up with the increasing demand. The ability of machine learning (ML) models to predict such decisions based on past cases under similar circumstances was assessed in some recent works. The dominant conclusion is that the prediction goal is achievable with high accuracy. Nevertheless, most of those works do not consider important (...) aspects for ML models that can impact performance and affect real-world usefulness, such as consistency, out-of-sample applicability, generality, and explainability preservation. To our knowledge, none considered all those aspects, and no previous study addressed the joint use of metadata and text-extracted variables to predict administrative decisions. We propose a predictive model that addresses the abovementioned concerns based on a two-stage cascade classifier. The model employs a first-stage prediction based on textual features extracted from the original documents and a second-stage classifier that includes proceedings’ metadata. The study was conducted using time-based cross-validation, built on data available before the predicted judgment. It provides predictions as soon as the decision date is scheduled and only considers the first document in each proceeding, along with the metadata recorded when the infringement is first registered. Finally, the proposed model provides local explainability by preserving visibility on the textual features and employing the SHapley Additive exPlanations (SHAP). Our findings suggest that this cascade approach surpasses the standalone stages and achieves relatively high Precision and Recall when both text and metadata are available while preserving real-world usefulness. With a weighted F1 score of 0.900, the results outperform the text-only baseline by 1.24% and the metadata-only baseline by 5.63%, with better discriminative properties evaluated by the receiver operating characteristic and precision-recall curves. (shrink)

This paper uses an improved deep learning algorithm to judge the rationality of the design of landscape image feature recognition. The preprocessing of the image is proposed to enhance the data. The deficiencies in landscape feature extraction are further addressed based on the new model. Then, the two-stage training method of the model is used to solve the problems of long training time and convergence difficulties in deep learning. Innovative methods for zoning and segmentation training of landscape (...) pattern features are proposed, which makes model training faster and generates more creative landscape patterns. Because of the impact of too many types of landscape elements in landscape images, traditional convolutional neural networks can no longer effectively solve this problem. On this basis, a fully convolutional neural network model is designed to perform semantic segmentation of landscape elements in landscape images. Through the method of deconvolution, the pixel-level semantic segmentation is realized. Compared with the 65% accuracy rate of the convolutional neural network, the fully convolutional neural network has an accuracy rate of 90.3% for the recognition of landscape elements. The method is effective, accurate, and intelligent for the classification of landscape element design, which better improves the accuracy of classification, greatly reduces the cost of landscape element design classification, and ensures that the technical method is feasible. This paper classifies landscape behavior based on this model for full convolutional neural network landscape images and demonstrates the effectiveness of using the model. In terms of landscape image processing, the image evaluation provides a certain basis. (shrink)

The intelligent transportation system under the big data environment is the development direction of the future transportation system. It effectively integrates advanced information technology, data communication transmission technology, electronic sensing technology, control technology, and computer technology and applies them to the entire ground transportation management system to establish a real-time, accurate, and efficient comprehensive transportation management system that works on a large scale and in all directions. Intelligent video analysis is an important part of smart transportation. In order to improve (...) the accuracy and time efficiency of video retrieval schemes and recognition schemes, this article firstly proposes a segmentation and key frame extraction method for video behavior recognition, using a multi-time scale dual-stream network to extract video features, improving the efficiency and efficiency of video behavior detection. On this basis, an improved algorithm for vehicle detection based on Faster R-CNN is proposed, and the Faster R-CNN network feature extraction layer is improved by using the principle of residual network, and a hole convolution is added to the network to filter out the redundant features of high-resolution video images to improve the problem of vehicle missed detection in the original algorithm. The experimental results show that the key frame extraction technology combined with the optimized Faster R-CNN algorithm model greatly improves the accuracy of detection and reduces the leakage. The detection rate is satisfactory. (shrink)

Transfer learning can improve the robustness of deep learning in the case of small samples. However, when the semantic difference between the source domain data and the target domain data is large, transfer learning easily introduces redundant features and leads to negative transfer. According the mechanism of the human brain focusing on effective features while ignoring redundant features in recognition tasks, a brain-like classification method based on adaptive feature matching dual-source domain heterogeneous transfer learning is proposed for the preoperative (...) aided diagnosis of lung granuloma and lung adenocarcinoma for patients with solitary pulmonary solid nodule in the case of small samples. The method includes two parts: (1) feature extraction and (2) feature classification. In the feature extraction part, first, By simulating the feature selection mechanism of the human brain in the process of drawing inferences about other cases from one instance, an adaptive selected-based dual-source domain feature matching network is proposed to determine the matching weight of each pair of feature maps and each pair of convolution layers between the two source networks and the target network, respectively. These two weights can, respectively, adaptive select the features in the source network that are conducive to the learning of the target task, and the destination of feature transfer to improve the robustness of the target network. Meanwhile, a target network based on diverse branch block is proposed, which made the target network have different receptive fields and complex paths to further improve the feature expression ability of the target network. Second, the convolution kernel of the target network is used as the feature extractor to extract features. In the feature classification part, an ensemble classifier based on sparse Bayesian extreme learning machine is proposed that can automatically decide how to combine the output of base classifiers to improve the classification performance. Finally, the experimental results (the AUCs were 0.9542 and 0.9356, respectively) on the data of two center data show that this method can provide a better diagnostic reference for doctors. (shrink)

Most current statistical natural language processing models use only local features so as to permit dynamic programming in inference, but this makes them unable to fully account for the long distance structure that is prevalent in language use. We show how to solve this dilemma with Gibbs sam- pling, a simple Monte Carlo method used to perform approximate inference in factored probabilistic models. By using simulated annealing in place of Viterbi decoding in sequence models such as HMMs, CMMs, and CRFs, (...) it is possible to incorporate non-local structure while preserving tractable inference. We use this technique to augment an existing CRF-based information extraction system with long-distance dependency models, enforcing label consistency and extraction template consistency constraints. This technique results in an error reduction of up to 9% over state-of-the-art systems on two established information extraction tasks. (shrink)

. Effective text region extraction and binarization of image embedded text documents on mobile devices having limited computational resources is an open research problem. In this paper, we present one such technique for preprocessing images captured with built-in cameras of handheld devices with an aim of developing an efficient Business Card Reader. At first, the card image is processed for isolating foreground components. These foreground components are classified as either text or non-text using different feature descriptors of texts (...) and images. The non-text components are removed and the textual ones are binarized with a fast adaptive algorithm. Specifically, we propose new techniques for foreground component isolation, text extraction and binarization of text regions from camera captured business card images. Experiments with business card images of various resolutions show that the present technique yields better accuracy and involves low computational overhead in comparison with the state-of-the-art. We achieve optimum text/non-text separation performance with images of resolution 800×600 pixels with an average recall rate of 93.90% and a precision rate of 96.84%. It involves a peak memory consumption of 0.68 MB and processing time of 0.102 seconds on a moderately powerful notebook, and 4 seconds of processing time on a PDA. (shrink)

We initially describe a feature-rich discriminative Conditional Random Field (CRF) model for Information Extraction in the workshop announcements domain, which offers good baseline performance in the PASCAL shared task. We then propose a method for leveraging domain knowledge in Information Extraction tasks, scoring candidate document labellings as one-value-per-field templates according to domain feasibility after generating sample labellings from a trained sequence classifier. Our relational models evaluate these templates according to our intuitions about agreement in the domain: workshop (...) acronyms should resemble their names, workshop dates occur after paper submission dates. These methods see a 5% f-score improvement in fields retrieved when sampling labellings from a Maximum-Entropy Markov Model, however we do not observe improvement over a CRF model. We discuss reasons for this, including the problem of recovering all field instances from a best template, and propose future work in adapting such a model to the CRF, a better standalone system. (shrink)

There are complex posture changes in dance movements, which lead to the low accuracy of dance movement recognition. And none of the current motion recognition uses the dancer’s attributes. The attribute feature of dancer is the important high-level semantic information in the action recognition. Therefore, a dance movement recognition algorithm based on feature expression and attribute mining is designed to learn the complicated and changeable dancer movements. Firstly, the original image information is compressed by the time-domain fusion module, (...) and the information of action and attitude can be expressed completely. Then, a two-way feature extraction network is designed, which extracts the details of the actions along the way and takes the sequence image as the input of the network. Then, in order to enhance the expression ability of attribute features, a multibranch spatial channel attention integration module based on an attention mechanism is designed to extract the features of each attribute. Finally, using the semantic inference and information transfer function of the graph convolution network, the relationship between attribute features and dancer features can be mined and deduced, and more expressive action features can be obtained; thus, high-performance dance motion recognition is realized. The test and analysis results on the data set show that the algorithm can recognize the dance movement and improve the accuracy of the dance movement recognition effectively, thus realizing the movement correction function of the dancer. (shrink)

This paper presents an analysis of the issue that as yet not been extensively researched in the doctrine of Lithuanian and foreign law: the issue of legal distinguishing between medicinal products and food supplements. In order to analyze the problems of theory and practice, the structure of the paper is divided into two parts. The first part concentrates on the main features of medicinal products and food supplements in accordance with the case law of the Court of Justice of the (...) European Union (CJEU). It is established that medicinal products must comply with one of these descriptions: be made and aimed towards healing and preventing sicknesses; cause serious threats to health or have health-hazardous secondary effects; and have very intense effects on physiological features of human organisms. Food supplements in theory should not have these features. However, sometimes in practice preparations identified as food supplements have large doses of certain medicinal substances or other ingredients (e.g. alcohol, plant extracts, vitamins, amino acids, and etc.) that present threats to human health. The practice of the CJEU shows that counterbalancing free movement and public health protection interests is problematic in this area. (shrink)

Electroencephalography and functional near-infrared spectroscopy have potentially complementary characteristics that reflect the electrical and hemodynamic characteristics of neural responses, so EEG-fNIRS-based hybrid brain-computer interface is the research hotspots in recent years. However, current studies lack a comprehensive systematic approach to properly fuse EEG and fNIRS data and exploit their complementary potential, which is critical for improving BCI performance. To address this issue, this study proposes a novel multimodal fusion framework based on multi-level progressive learning with multi-domain features. The framework consists (...) of a multi-domain feature extraction process for EEG and fNIRS, a feature selection process based on atomic search optimization, and a multi-domain feature fusion process based on multi-level progressive machine learning. The proposed method was validated on EEG-fNIRS-based motor imagery and mental arithmetic tasks involving 29 subjects, and the experimental results show that multi-domain features provide better classification performance than single-domain features, and multi-modality provides better classification performance than single-modality. Furthermore, the experimental results and comparison with other methods demonstrated the effectiveness and superiority of the proposed method in EEG and fNIRS information fusion, it can achieve an average classification accuracy of 96.74% in the MI task and 98.42% in the MA task. Our proposed method may provide a general framework for future fusion processing of multimodal brain signals based on EEG-fNIRS. (shrink)

Several studies have shown that our visual system may construct a “summary statistical representation” over groups of visual objects. Although there is a general understanding that human observers can accurately represent sets of a variety of features, many questions on how summary statistics, such as an average, are computed remain unanswered. This study investigated sampling properties of visual information used by human observers to extract two types of summary statistics of item sets, average and variance. We presented three models of (...) ideal observers to extract the summary statistics: a global sampling model without sampling noise, global sampling model with sampling noise, and limited sampling model. We compared the performance of an ideal observer of each model with that of human observers using statistical efficiency analysis. Results suggest that summary statistics of items in a set may be computed without representing individual items, which makes it possible to discard the limited sampling account. Moreover, the extraction of summary statistics may not necessarily require the representation of individual objects with focused attention when the sets of items are larger than 4. (shrink)

As one of the most widely used methods in deep learning technology, convolutional neural networks have powerful feature extraction capabilities and nonlinear data fitting capabilities. However, the convolutional neural network method still has disadvantages such as complex network model, too long training time and excessive consumption of computing resources, slow convergence speed, network overfitting, and classification accuracy that needs to be improved. Therefore, this article proposes a dense convolutional neural network classification algorithm based on texture features for images (...) in virtual reality videos. First, the texture feature of the image is introduced as a priori information to reflect the spatial relationship between pixels and the unique characteristics of different types of ground features. Second, the grey level cooccurrence matrix is used to extract the grey level correlation features of the image in space. Then, Gauss Markov Random Field is used to establish the statistical correlation characteristics between neighbouring pixels, and the extracted GLCM-GMRF texture feature and image intensity vector are combined. Finally, based on DenseNet, an improved shallow layer dense convolutional neural network is proposed, which can compress network parameters and improve the feature extraction ability of the network. The experimental results show that compared with the current classification method, this method can effectively suppress the influence of coherent speckle noise and obtain better classification results. (shrink)

Audio features such as inharmonicity, noisiness, and spectral roll-off have been identified as correlates of “noisy” sounds. However, such features are likely involved in the experience of multiple semantic timbre categories of varied meaning and valence. This paper examines the relationships of stimulus properties and audio features with the semantic timbre categories raspy/grainy/rough, harsh/noisy, and airy/breathy. Participants rated a random subset of 52 stimuli from a set of 156 approximately 2-s orchestral instrument sounds representing varied instrument families, registers, and both (...) traditional and extended playing techniques. Stimuli were rated on the three semantic categories of interest, as well as on perceived playing exertion and emotional valence. Correlational analyses demonstrated a strong negative relationship between positive valence and perceived physical exertion. Exploratory linear mixed models revealed significant effects of extended technique and pitch register on valence, the perception of physical exertion, raspy/grainy/rough, and harsh/noisy. Instrument family was significantly related to ratings of airy/breathy. With an updated version of the Timbre Toolbox, we used 44 summary audio features, extracted from the stimuli using spectral and harmonic representations, as input for various models built to predict mean semantic ratings for each sound on the three semantic categories, on perceived exertion, and on valence. Random Forest models predicting semantic ratings from audio features outperformed Partial Least-Squares Regression models, consistent with previous results suggesting that non-linear methods are advantageous in timbre semantic predictions using audio features. Relative Variable Importance measures from the models among the three semantic categories demonstrate that although these related semantic categories are associated in part with overlapping features, they can be differentiated through individual patterns of audio feature relationships. (shrink)

Humans naturally perceive visual patterns in a global manner and are remarkably capable of extracting object shapes based on properties such as proximity, closure, symmetry, and good continuation. Notwithstanding the role of these properties in perceptual grouping, studies highlighted differences in disembedding performance across individuals, which are summarized by the field dependence dimension. Evidence suggests that age and educational attainment explain part of this variability, whereas the role of sex is still highly debated. Also, which stimulus features primarily influence inter-individual (...) variations in perceptual grouping has still to be fully determined. Building upon these premises, we assessed the role of age, education level, and sex on performance at the Leuven Embedded Figure Test—a proxy of disembedding abilities—in 391 cisgender individuals. We also investigated to what extent shape symmetry, closure, complexity, and continuation relate to task accuracy. Overall, target asymmetry, closure, and good continuation with the embedding context increase task difficulty. Simpler shapes are more difficult to detect than those with more lines, yet context complexity impairs the recognition of complex targets to a greater extent. Concerning demographic data, we confirm that age and educational attainment are significantly associated with disembedding abilities and reveal a perceptual advantage in males. In summary, our study further highlights the role of shape properties in disembedding performance and unveils sex differences not reported so far. (shrink)

Multi-person pose estimation is a challenging problem. Bottom-up methods have been greatly studied because the prediction speed of top-down methods is related to the number of people in the input image, making these methods difficult to apply in real-time environments. To solve the problems of scale sensitivity and quantization error in bottom-up methods, it is necessary to have a model that can predict multi-scale keypoints and refine quantization error. To achieve this, we propose context feature and refined network for (...) multi-person pose estimation (CRNet), which can effectively solve the problems of scale sensitivity and quantization error in bottom-up methods. We use a multi-scale feature pyramid and context feature to achieve scale invariance of the network. We extract global and local features and then fuse them by attentional feature fusion (AFF) to obtain context feature that adapt to multi-scale keypoints. In addition, we propose an efficient refined network to solve the problem of quantization error and use multi-resolution supervised learning to further improve the prediction accuracy of CRNet. Comprehensive experiments are conducted on two benchmarks: COCO and MPII datasets. The average precision of CRNet reached 72.1 and 80.2%, respectively, surpassing most state-of-the-art methods. (shrink)

A brain tumor is an abnormal mass or growth of a cell that leads to certain death, and this is still a challenging task in clinical practice. Early and correct diagnosis of this type of cancer is very important for the treatment process. For this reason, this study aimed to develop computer-aided systems for the diagnosis of brain tumors. In this research, we proposed three different end-to-end deep learning approaches for analyzing effects of local and deep features for brain MRI (...) images anomaly detection. The first proposed system is Directional Bit-Planes Deep Autoencoder which extracts and learns local and direction features. The DBP-DAE by decomposition of a local binary pattern into eight bit-planes extracts are directional and inherent local-structure features from the input image and learns robust feature for classification purposes. The second one is a Dilated Separable Residual Convolutional Network which extracts high and low-level features. The main advantage of this approach is that it is robust and shows stable results regardless to size of image database and to solve overfitting problems. To explore the effects of mixture of local and deep extracted feature on accuracy of classification of brain anomaly, a multibranch convolutional neural network approach is proposed. This approach is designed according to combination of DBP-DAE and DSRCN in an end-to-end manner. Extensive experiments conducted based on brain tumor in MRI image public access databases and achieves significant results compared to state-of-the-art algorithms. In addition, we discussed the effectiveness and applicability of CNNs with a variety of different features and architectures for brain abnormalities such as Alzheimer’s. (shrink)

Numerous experiments have examined whether moving stimuli capture spatial attention but none have sought to determine whether visual features of looming and receding objects are extracted in a capacity-free manner. The current experiment used the task-choice procedure originated by Besner and Care to examine this possibility. Stimuli were presented in 3D space by manipulating retinal disparity. Results indicate that features of an object are extracted in a capacity-free manner for both looming and receding objects for participants who consciously perceive motion (...) but not for participants who do not consciously perceive motion. These results suggest that the cognitive system is biased to process potentially animate objects, perhaps because of the evolutionary advantage this cognitive ability may provide. (shrink)