In step with rapid advancements in computer vision, vehicle classification demonstrates a considerable potential to reshape intelligent transportation systems. In the last couple of decades, image processing and pattern recognition-based vehicle classification systems have been used to improve the effectiveness of automated highway toll collection and traffic monitoring systems. However, these methods are trained on limited handcrafted features extracted from small datasets, which do not cater the real-time road traffic conditions. Deep learning-based classification systems have been proposed to incorporate the (...) above-mentioned issues in traditional methods. However, convolutional neural networks require piles of data including noise, weather, and illumination factors to ensure robustness in real-time applications. Moreover, there is no generalized dataset available to validate the efficacy of vehicle classification systems. To overcome these issues, we propose a convolutional neural network-based vehicle classification system to improve robustness of vehicle classification in real-time applications. We present a vehicle dataset comprising of 10,000 images categorized into six-common vehicle classes considering adverse illuminous conditions to achieve robustness in real-time vehicle classification systems. Initially, pretrained AlexNet, GoogleNet, Inception-v3, VGG, and ResNet are fine-tuned on self-constructed vehicle dataset to evaluate their performance in terms of accuracy and convergence. Based on better performance, ResNet architecture is further improved by adding a new classification block in the network. To ensure generalization, we fine-tuned the network on the public VeRi dataset containing 50,000 images, which have been categorized into six vehicle classes. Finally, a comparison study has been carried out between the proposed and existing vehicle classification methods to evaluate the effectiveness of the proposed vehicle classification system. Consequently, our proposed system achieved 99.68%, 99.65%, and 99.56% accuracy, precision, and F1-score on our self-constructed dataset. (shrink)

Given the extent of using deep convolutional neural networks to model the mechanism of object recognition, it becomes important to analyse the evidence of their similarity and the explanatory potential of these models. I focus on one frequent method of their comparison—representational similarity analysis, and I argue, first, that it underdetermines these models as how-actually mechanistic explanations. This happens because different similarity measures in this framework pick out different mechanisms across DCNNs and the brain in order to correspond them, and (...) there is no arbitration between them in terms of relevance for object recognition. Second, the reason similarity measures are underdetermining to a large degree stems from the highly idealised nature of these models, which undermines their status as how-possibly mechanistic explanatory models of object recognition as well. Thus, building models with more theoretical consideration and choosing relevant similarity measures may bring us closer to the goal of mechanistic explanation. (shrink)

According to Thagard and Stewart :1–33, 2011), creativity results from the combination of neural representations, and combination results from convolution, an operation on vectors defined in the holographic reduced representation framework. They use these ideas to understand creativity as it occurs in many domains, and in particular in science. We argue that, because of its algebraic properties, convolution alone is ill-suited to the role proposed by Thagard and Stewart. The semantic pointer concept allows us to see how we (...) can apply the full range of HRR operations while keeping the modal representations so central to Thagard and Stewart’s theory. By adding another combination operation and using semantic pointers as the combinatorial basis, this modified version overcomes the limitations of the original theory and perhaps helps us explain aspects of creativity not covered by the original theory. While a priori reasons cast doubts on the use of HRR operations with modal representations :5039–5054, 1987) such as semantic pointers, recent models point in the other direction, allowing us to be optimistic about the success of the revised version. (shrink)

According to Thagard and Stewart :1–33, 2011), creativity results from the combination of neural representations, and combination results from convolution, an operation on vectors defined in the holographic reduced representation framework. They use these ideas to understand creativity as it occurs in many domains, and in particular in science. We argue that, because of its algebraic properties, convolution alone is ill-suited to the role proposed by Thagard and Stewart. The semantic pointer concept allows us to see how we (...) can apply the full range of HRR operations while keeping the modal representations so central to Thagard and Stewart’s theory. By adding another combination operation and using semantic pointers as the combinatorial basis, this modified version overcomes the limitations of the original theory and perhaps helps us explain aspects of creativity not covered by the original theory. While a priori reasons cast doubts on the use of HRR operations with modal representations :5039–5054, 1987) such as semantic pointers, recent models point in the other direction, allowing us to be optimistic about the success of the revised version. (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)

The skin, uniquely positioned at the interface between the human body and the external world, plays a multifaceted immunologic role in human life. In medical practice, early accurate detection of all types of skin tumours is essential to guide appropriate management and improve patients’ survival. The most important issue is to differentiate between malignant skin tumours and benign lesions. The aim of this research is the classification of skin tumours by analysing medical skin tumour dermoscopy images. This paper is focused (...) on a new strategy based on deep convolutional neural networks which have recently shown a state-of-the-art performance to define strategy to automatic classification for skin tumour images. The proposed system is tested on well-known HAM10000 data set. For experimental results, verification is performed and the results are compared with similar researches. (shrink)

Cultural similarities and differences in facial expressions have been a controversial issue in the field of facial communications. A key step in addressing the debate regarding the cultural dependency of emotional expression is to characterize the visual features of specific facial expressions in individual cultures. Here we developed an image analysis framework for this purpose using convolutional neural networks that through training learned visual features critical for classification. We analyzed photographs of facial expressions derived from two databases, each developed in (...) a different country, in which corresponding emotion labels were available. While the CNNs reached high rates of correct results that were far above chance after training with each database, they showed many misclassifications when they analyzed faces from the database that was not used for training. These results suggest that facial features useful for classifying facial expressions differed between the databases. The selectivity of computational units in the CNNs to action units of the face varied across the facial expressions. Importantly, the AU selectivity often differed drastically between the CNNs trained with the different databases. Similarity and dissimilarity of these tuning profiles partly explained the pattern of misclassifications, suggesting that the AUs are important for characterizing the facial features and differ between the two countries. The AU tuning profiles, especially those reduced by principal component analysis, are compact summaries useful for comparisons across different databases, and thus might advance our understanding of universality vs. specificity of facial expressions across cultures. (shrink)

In artificial intelligence, recent research has demonstrated the remarkable potential of Deep Convolutional Neural Networks (DCNNs), which seem to exceed state-of-the-art performance in new domains weekly, especially on the sorts of very difficult perceptual discrimination tasks that skeptics thought would remain beyond the reach of artificial intelligence. However, it has proven difficult to explain why DCNNs perform so well. In philosophy of mind, empiricists have long suggested that complex cognition is based on information derived from sensory experience, often appealing to (...) a faculty of abstraction. Rationalists have frequently complained, however, that empiricists never adequately explained how this faculty of abstraction actually works. In this paper, I tie these two questions together, to the mutual benefit of both disciplines. I argue that the architectural features that distinguish DCNNs from earlier neural networks allow them to implement a form of hierarchical processing that I call “transformational abstraction”. Transformational abstraction iteratively converts sensory-based representations of category exemplars into new formats that are increasingly tolerant to “nuisance variation” in input. Reflecting upon the way that DCNNs leverage a combination of linear and non-linear processing to efficiently accomplish this feat allows us to understand how the brain is capable of bi-directional travel between exemplars and abstractions, addressing longstanding problems in empiricist philosophy of mind. I end by considering the prospects for future research on DCNNs, arguing that rather than simply implementing 80s connectionism with more brute-force computation, transformational abstraction counts as a qualitatively distinct form of processing ripe with philosophical and psychological significance, because it is significantly better suited to depict the generic mechanism responsible for this important kind of psychological processing in the brain. (shrink)

Timely detection and efficient recognition of fault are challenging for the bogie of high-speed train, owing to the fact that different types of fault signals have similar characteristics in the same frequency range. Notice that convolutional neural networks are powerful in extracting high-level local features and that recurrent neural networks are capable of learning long-term context dependencies in vibration signals. In this paper, by combining CNN and RNN, a so-called convolutional recurrent neural network is proposed to diagnose various faults of (...) the HST bogie, where the capabilities of CNN and RNN are inherited simultaneously. Within the novel architecture, the proposed CRNN first filters out the features from the original data through convolutional layers. Then, four recurrent layers with simple recurrent cell are used to model the context information in the extracted features. By comparing the performance of the presented CRNN with CNN, RNN, and ensemble learning, experimental results show that CRNN achieves not only the best performance with accuracy of 97.8% but also the least time spent in training model. (shrink)

Convolutional neural networks, which automatically learn features from raw data to approximate functions, are being increasingly applied to the end-to-end analysis of electroencephalographic signals, especially for decoding brain states in brain-computer interfaces. Nevertheless, CNNs introduce a large number of trainable parameters, may require long training times, and lack in interpretability of learned features. The aim of this study is to propose a CNN design for P300 decoding with emphasis on its lightweight design while guaranteeing high performance, on the effects of (...) different training strategies, and on the use of post-hoc techniques to explain network decisions. The proposed design, named MS-EEGNet, learned temporal features in two different timescales in an efficient and optimized way, and was validated on three P300 datasets. The CNN was trained using different strategies and was compared with several state-of-the-art algorithms. Furthermore, variants of the baseline MS-EEGNet were analyzed to evaluate the impact of different hyper-parameters on performance. Lastly, saliency maps were used to derive representations of the relevant spatio-temporal features that drove CNN decisions. MS-EEGNet was the lightest CNN compared with the tested SOA CNNs, despite its multiple timescales, and significantly outperformed the SOA algorithms. Post-hoc hyper-parameter analysis confirmed the benefits of the innovative aspects of MS-EEGNet. Furthermore, MS-EEGNet did benefit from transfer learning, especially using a low number of training examples, suggesting that the proposed approach could be used in BCIs to accurately decode the P300 event while reducing calibration times. Representations derived from the saliency maps matched the P300 spatio-temporal distribution, further validating the proposed decoding approach. This study, by specifically addressing the aspects of lightweight design, transfer learning, and interpretability, can contribute to advance the development of deep learning algorithms for P300-based BCIs. (shrink)

The investigation of visual categorization has recently been aided by the introduction of deep convolutional neural networks (CNNs), which achieve unprecedented accuracy in picture classification after extensive training. Even if the architecture of CNNs is inspired by the organization of the visual brain, the similarity between CNN and human visual processing remains unclear. Here, we investigated this issue by engaging humans and CNNs in a two‐class visual categorization task. To this end, pictures containing animals or vehicles were modified to contain (...) only low/high spatial frequency (HSF) information, or were scrambled in the phase of the spatial frequency spectrum. For all types of degradation, accuracy increased as degradation was reduced for both humans and CNNs; however, the thresholds for accurate categorization varied between humans and CNNs. More remarkable differences were observed for HSF information compared to the other two types of degradation, both in terms of overall accuracy and image‐level agreement between humans and CNNs. The difficulty with which the CNNs were shown to categorize high‐passed natural scenes was reduced by picture whitening, a procedure which is inspired by how visual systems process natural images. The results are discussed concerning the adaptation to regularities in the visual environment (scene statistics); if the visual characteristics of the environment are not learned by CNNs, their visual categorization may depend only on a subset of the visual information on which humans rely, for example, on low spatial frequency information. (shrink)

Real-time smoke detection is of great significance for early warning of fire, which can avoid the serious loss caused by fire. Detecting smoke in actual scenes is still a challenging task due to large variance of smoke color, texture, and shapes. Moreover, the smoke detection in the actual scene is faced with the difficulties in data collection and insufficient smoke datasets, and the smoke morphology is susceptible to environmental influences. To improve the performance of smoke detection and solve the problem (...) of too few datasets in real scenes, this paper proposes a model that combines a deep convolutional generative adversarial network and a convolutional neural network to extract smoke features and detection. The vibe algorithm was used to collect smoke and nonsmoke images in the dynamic scene and deep convolutional generative adversarial network used these images to generate images that are as realistic as possible. Besides, we designed an improved convolutional neural network model for extracting smoke features and smoke detection. The experimental results show that the method has a good detection performance on the smoke generated in the actual scenes and effectively reduces the false alarm rate. (shrink)

Augmented reality-based brain–computer interface (AR–BCI) has a low signal-to-noise ratio (SNR) and high real-time requirements. Classical machine learning algorithms that improve the recognition accuracy through multiple averaging significantly affect the information transfer rate (ITR) of the AR–SSVEP system. In this study, a fast recognition method based on a separable convolutional neural network (SepCNN) was developed for an AR-based P300 component (AR–P300). SepCNN achieved single extraction of AR–P300 features and improved the recognition speed. A nine-target AR–P300 single-stimulus paradigm was designed to (...) be administered with AR holographic glasses to verify the effectiveness of SepCNN. Compared with four classical algorithms, SepCNN significantly improved the average target recognition accuracy (81.1%) and information transmission rate (57.90 bits/min) of AR–P300 single extraction. SepCNN with single extraction also attained better results than classical algorithms with multiple averaging. (shrink)

This paper draws together empirical findings from our study of hackathons in the UK with literature on big data through three interconnected frameworks: data as discourse, data as datalogical and data as materiality. We suggest not only that hackathons resonate the wider socio-technical and political constructions of data that are currently enacted in policy, education and the corporate sector, but also that an investigation of hackathons reveals the extent to which ‘data’ operates as a powerful discursive tool; how the discourses (...) of data mask and reveal a series of tropes pertaining to data; that the politics of data are routinely and simultaneously obscured and claimed with serious implications for expertise and knowledge; and that ultimately, and for the vast majority of hackathons we have attended, the discursive and material constructions of data serve to underpin rather than challenge existing power relations and politics. (shrink)

The target and background will change continuously in the long-term tracking process, which brings great challenges to the accurate prediction of targets. The correlation filter algorithm based on manual features is difficult to meet the actual needs due to its limited feature representation ability. Thus, to improve the tracking performance and robustness, an improved hierarchical convolutional features model is proposed into a correlation filter framework for visual object tracking. First, the objective function is designed by lasso regression modeling, and a (...) sparse, time-series low-rank filter is learned to increase the interpretability of the model. Second, the features of the last layer and the second pool layer of the convolutional neural network are extracted to realize the target position prediction from coarse to fine. In addition, using the filters learned from the first frame and the current frame to calculate the response maps, respectively, the target position is obtained by finding the maximum response value in the response map. The filter model is updated only when these two maximum responses meet the threshold condition. The proposed tracker is evaluated by simulation analysis on TC-128/OTB2015 benchmarks including more than 100 video sequences. Extensive experiments demonstrate that the proposed tracker achieves competitive performance against state-of-the-art trackers. The distance precision rate and overlap success rate of the proposed algorithm on OTB2015 are 0.829 and 0.695, respectively. The proposed algorithm effectively solves the long-term object tracking problem in complex scenes. (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)

Grodzinsky's localization claim can be questioned on empirical grounds. The Trace Deletion Hypothesis fails to account for a number of comprehension facts in Broca's aphasia and conduction aphasics show similar comprehension patterns. Frontoparietal systems are recruited during sentence comprehension only under conditions of increased processing load and/or attentional demands.

Focusing principally on the once-feted philosophical movement of object-oriented ontology (OOO), this article examines the ways in which this movement fits into a broader “speculative turn,” which seeks to reverse the purportedly wrongheaded emphasis of post-Kantian critical philosophy upon the finitude of the subject and to once again unleash the fecund potentialities of speculative thought. Identifying several incongruities and tensions that traverse this project, it is argued that OOO exemplifies the difficulties faced when attempting to articulate a decidedly pre-critical metaphysics.

The purpose is to improve the teaching and learning efficiency of college Innovation and Entrepreneurship Education. Firstly, from the perspective of aesthetic education, this work designs the teacher and student sides of the Computer-aided Instruction system. Secondly, the CAI model is implemented based on the weight sharing and local perception of the Convolutional Neural Network. Finally, the performance of the CNN-based CAI model is tested. Meanwhile, it analyses students’ IEE experience under the proposed CAI model through a case study of (...) Music Majors from Xi’an Conservatory of Music. The experimental data show that the CNN-based CAI model can respond quickly and stably when users access different functional modules, such as webpage browsing. The proposed CAI model increases students’ entrepreneurial interest, skills, and knowledge by 55.62, 57.32, and 72.12%, respectively. Students’ entrepreneurial practice ability has been improved by over 50.00%; such an increase in entrepreneurial practice ability has also shown individual differences. Thus, the proposed Music Majors-oriented IEE-infiltrated CAI model based on CNN improves students’ entrepreneurial practice ability and reflects the positive experience of Music Majors on IEE. The finding provides references for the step-by-step identification of the CNN-based CAI model and has certain guiding significance for analyzing the effect of college IEE. (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)

Faces provide information about a person's identity, as well as their sex, age, and ethnicity. People also infer social and personality traits from the face — judgments that can have important societal and personal consequences. In recent years, deep convolutional neural networks (DCNNs) have proven adept at representing the identity of a face from images that vary widely in viewpoint, illumination, expression, and appearance. These algorithms are modeled on the primate visual cortex and consist of multiple processing layers of simulated (...) neurons. Here, we examined whether a DCNN trained for face identification also retains a representation of the information in faces that supports social‐trait inferences. Participants rated male and female faces on a diverse set of 18 personality traits. Linear classifiers were trained with cross validation to predict human‐assigned trait ratings from the 512 dimensional representations of faces that emerged at the top‐layer of a DCNN trained for face identification. The network was trained with 494,414 images of 10,575 identities and consisted of seven layers and 19.8 million parameters. The top‐level DCNN features produced by the network predicted the human‐assigned social trait profiles with good accuracy. Human‐assigned ratings for the individual traits were also predicted accurately. We conclude that the face representations that emerge from DCNNs retain facial information that goes beyond the strict limits of their training. (shrink)

With the accelerated evolution of the internet in the form of web-sites, social networks, microblogs, and online portals, a large number of reviews, opinions, recommendations, ratings, and feedback are generated by writers or users. This user-generated sentiment content can be about books, people, hotels, products, research, events, etc. These sentiments become very beneficial for businesses, governments, and individuals. While this content is meant to be useful, a bulk of this writer-generated content requires using text mining techniques and sentiment analysis. However, (...) there are several challenges facing the sentiment analysis and evaluation process. These challenges become obstacles in analyzing the accurate meaning of sentiments and detecting suitable sentiment polarity specifically in the Bangla language. Sentiment analysis is the practice of applying natural language processing and text analysis techniques to identify and extract subjective information from text. This paper presents how the attention mechanism could be incorporated effectively and efficiently in analyzing the Bangla sentiment or opinion. (shrink)

This article examines motifs of falling, recoiling, and turning across Chester Himes’ oeuvre as figurations of Black susceptibility to racial violence. These images reference and reconstruct an event from Himes’ early adulthood: his catastrophic fall down an elevator shaft. Taking a psychoanalytically oriented approach, I analyze the metonymic connections between these motifs, rather than reading them in their chronological order, using Jean Laplanche’s theory of après-coup. I argue that the recursive quality of these images in Himes’ work is not merely (...) an unconscious repetition or conscious working through of a traumatic biographical event but part of an endeavor to imagine different ways to inhabit and survive the structural trauma of Jim Crow America. (shrink)

In the field of natural language processing, the task of sensitive information detection refers to the procedure of identifying sensitive words for given documents. The majority of existing detection methods are based on the sensitive-word tree, which is usually constructed via the common prefixes of different sensitive words from the given corpus. Yet, these traditional methods suffer from a couple of drawbacks, such as poor generalization and low efficiency. For improvement purposes, this paper proposes a novel self-attention-based detection algorithm using (...) the implementation of graph convolutional network. The main contribution is twofold. Firstly, we consider a weighted GCN to better encode word pairs from the given documents and corpus. Secondly, a simple, yet effective, attention mechanism is introduced to further integrate the interaction among candidate words and corpus. Experimental results from the benchmarking dataset of THUC news demonstrate a promising detection performance, compared to existing work. (shrink)

In recent years, more and more people are applying Convolutional Neural Networks to the study of sound signals. The main reason is the translational invariance of convolution in time and space. Thereby the diversity of the sound signal can be overcome. However, in terms of sound direction recognition, there are also problems such as a microphone matrix being too large, and feature selection. This paper proposes a sound direction recognition using a simulated human head with microphones at both ears. (...) Theoretically, the two microphones cannot distinguish the front and rear directions. However, we use the original data of the two channels as the input of the convolutional neural network, and the resolution effect can reach more than 0.9. For comparison, we also chose the delay feature (GCC) for sound direction recognition. Finally, we also conducted experiments that used probability distributions to identify more directions. (shrink)

This study decodes consumers' preference levels using a convolutional neural network in neuromarketing. The classification accuracy in neuromarketing is a critical factor in evaluating the intentions of the consumers. Functional near-infrared spectroscopy is utilized as a neuroimaging modality to measure the cerebral hemodynamic responses. In this study, a specific decoding structure, called CNN-based fNIRS-data analysis, was designed to achieve a high classification accuracy. Compared to other methods, the automated characteristics, constant training of the dataset, and learning efficiency of the proposed (...) method are the main advantages. The experimental procedure required eight healthy participants to view commercial advertisement videos of different durations. The cerebral hemodynamic responses of the participants were measured. To compare the preference classification performances, CNN was utilized to extract the most common features, including the mean, peak, variance, kurtosis, and skewness. Considering three video durations, the average classification accuracies of 15, 30, and 60 s videos were 84.3, 87.9, and 86.4%, respectively. Among them, the classification accuracy of 87.9% for 30 s videos was the highest. The average classification accuracies of three preferences in females and males were 86.2 and 86.3%, respectively, showing no difference in each group. By comparing the classification performances in three different combinations between two groups, male participants were observed to have targeted preferences for commercial advertising, and the classification performance 88.4% between “like” vs. “dislike” out of three categories was the highest. Finally, pairwise classification performance are shown as follows: For female, 86.1%, 87.4%, 85.2%, and for male 85.7, 88.4, 85.1%, respectively. (shrink)

Early and accurate diagnosis of glaucoma is critical for avoiding human vision deterioration and preventing blindness. A deep-neural-network model has been developed for the diagnosis of glaucoma based on Heidelberg retina tomography, called “Seeking Common Features and Reserving Differences Net” to make full use of the HRT data. In this work, the proposed SCRD-Net model achieved an area under the curve of 94.0%. For the two HRT image modalities, the model sensitivities were 91.2% and 78.3% at specificities of 0.85 and (...) 0.95, respectively. These results demonstrate a significant improvement over earlier results. In addition, we visualized the network outputs to develop an interpretation of the learned mechanism for discriminating glaucoma and normal images. Thus, the SCRD-Net can be an effective diagnostic indicator of glaucoma during clinical screening. To facilitate SCRD-Net utilization by the scientific community, the code implementation will be made publicly available. (shrink)

Most binaural speech source localization models perform poorly in unprecedentedly noisy and reverberant situations. Here, this issue is approached by modelling a multiscale dilated convolutional neural network. The time-related crosscorrelation function and energy-related interaural level differences are preprocessed in separate branches of dilated convolutional network. The multiscale dilated CNN can encode discriminative representations for CCF and ILD, respectively. After encoding, the individual interaural representations are fused to map source direction. Furthermore, in order to improve the parameter adaptation, a novel semiadaptive (...) entropy is proposed to train the network under directional constraints. Experimental results show the proposed method can adaptively locate speech sources in simulated noisy and reverberant environments. (shrink)

Image-based crowd counting has extremely important applications in public safety issues. Most of the previous studies focused on extremely dense crowds. However, as the number of webcams increases, a crowd with extremely high density can obtain less error by summing the images of multiple close-range webcams, but there are still some problems such as heavy occlusions and large-scale variation. To solve the above problems, this paper proposes a new type of multibranch neural network with a compensator, in which features are (...) extracted through multibranch subnetworks of different scales. The weights between the branches are adjusted by the compensator, and the captured features are distinguished among different branches. To avoid learning nearly the same features in each branch and reducing the training deviation, the dataset is labeled with head scale, and the adaptive grading loss function is used to calculate the estimated loss of the subregions. The experimental results show that the accuracy of the network proposed in this paper is about 10% higher than that of the comparison network. (shrink)