Search

In this paper, a three-step model based on the integration of Deep Neural Networks (DNN) and Decision Models is introduced for image classification which is inspired by the human visual system. To make a decision about an object, many actions should be done in a hierarchical process in the brain. First, the retina receives visual stimuli and transfers them to the visual cortex in the brain. The information extracted in the visual cortex, is accumulated over time to select an appropriate response. Many of the current decision-making models do not show how each image is converted into useful information for the decision model. Some models have used neural networks to convert each image into the information needed in the decision-making model; however, the role of the retina is ignored among these models. In this paper, a combination of retina inspired filters, CNN-based description and accumulator-based decision model is used to classify images. This model's structure resembles the human brain due to the usage of the DoG filter bank as retina inspired filter in the first stage of it. This model shows a significant improvement in accuracy in comparison to other models; furthermore, its performance is acceptable even with the small sample training set. [ABSTRACT FROM AUTHOR]

Pruning is a useful technique for decreasing the memory consumption and floating point operations (FLOPs) of deep convolutional neural network (CNN) models. Nevertheless, at modest pruning levels, current structured pruning approaches often lead to considerable declines in accuracy. Furthermore, existing approaches often treat pruning rates as super parameters, neglecting the sensitivity of different convolution layers. In this study, we propose a novel sensitivity-based method for channel pruning that utilizes second-order sensitivity as a criterion. The essential concept is to prune insensitive filters while retaining sensitive ones. We quantify the sensitivity of the filter using the sum of the sensitivities of all weights in the filter, rather than the magnitude-based metric frequently applied in the literature. Furthermore, a layer sensitivity approach based on the Hessian eigenvalues of each layer is introduced into the process of automatically choosing the most appropriate pruning rate for each layer. Experiments on a variety of modern CNN architectures demonstrate that we can considerably enhance the pruning rate while sacrificing a small amount of accuracy, resulting in a reduction of more than 60% in FLOPs on CIFAR-10. Notably, on ImageNet, pruning based on ResNet50 decreased the FLOPs by 56.3% while losing only 0.92% of accuracy. [ABSTRACT FROM AUTHOR]

In personal identity recognition systems, detecting a person's age, gender, and language using voice signal characteristics is a crucial issue, especially because of the importance of security considerations. Age, gender, and language classification problems are important in signal processing because they are used to analyze and understand human behavior, interactions, and preferences. This can be especially useful in the fields of human-computer interaction, psychology, and social science research. In this paper, a new system for detecting a speaker's age, gender, and language based on deep learning models is presented. Deep learning models have shown great efficacy in various fields of signal processing. For this paper, a range of deep models were tested, including convolutional neural networks (CNNs), recurrent neural network (RNN), and a fine-tuning ResNet34 architecture. Additionally, techniques such as transfer learning were applied to improve the efficiency of the proposed system. The input voice signals are preprocessed by applying the spectro-temporal transform to obtain additional features that can be fed to the ResNet34 model, which is designed specifically for the task of voice signal processing. The dataset used in this paper was sourced from the Mozilla common voice initiative, which is dedicated to advancing speech recognition and language identification technologies. The performance of the proposed algorithm was evaluated in the presence of Gaussian noise to determine its robustness. The experimental results demonstrated that the proposed algorithm significantly outperformed basic algorithms and other deep neural networks in terms of age and gender recognition from voice signals. [ABSTRACT FROM AUTHOR]

• Neurobiological-inspired CNN models accurately predict sound location. • Models of binaural integration in human brainstem simulate human sound localization. • Objective functions minimizing angular or Euclidean distance lead to distinct errors. • Feature representations of shallow layers resemble brainstem neuronal spatial tuning. • Neurobiological-inspired CNN models generate testable predictions for neuroscience. The human brain effortlessly solves the complex computational task of sound localization using a mixture of spatial cues. How the brain performs this task in naturalistic listening environments (e.g. with reverberation) is not well understood. In the present paper, we build on the success of deep neural networks at solving complex and high-dimensional problems [1] to develop goal-driven, neurobiological-inspired convolutional neural network (CNN) models of human spatial hearing. After training, we visualize and quantify feature representations in intermediate layers to gain insights into the representational mechanisms underlying sound location encoding in CNNs. Our results show that neurobiological-inspired CNN models trained on real-life sounds spatialized with human binaural hearing characteristics can accurately predict sound location in the horizontal plane. CNN localization acuity across the azimuth resembles human sound localization acuity, but CNN models outperform human sound localization in the back. Training models with different objective functions - that is, minimizing either Euclidean or angular distance - modulates localization acuity in particular ways. Moreover, different implementations of binaural integration result in unique patterns of localization errors that resemble behavioral observations in humans. Finally, feature representations reveal a gradient of spatial selectivity across network layers, starting with broad spatial representations in early layers and progressing to sparse, highly selective spatial representations in deeper layers. In sum, our results show that neurobiological-inspired CNNs are a valid approach to modeling human spatial hearing. This work paves the way for future studies combining neural network models with empirical measurements of neural activity to unravel the complex computational mechanisms underlying neural sound location encoding in the human auditory pathway. [ABSTRACT FROM AUTHOR]

• Developing a new recursive convolutional layer (RCL) able to stop the iteration when the hidden state stabilize. • First network to have different depth for different batch of samples. • A comprehensive study on the best locations and numbers of RCL in a multilayer network called C-FRPN. To overcome problems with the design of large networks, particularly with respect to the depth of the network, this paper presents a new model of convolutional neural networks (CNN) which features fully recursive convolutional layers (RCLs). An RCL is a generalization of the classic one-stage feedforward convolutional layer (CL) to fully direct feedback connections between the outputs of the CL and its inputs. A traditional deep CNN consisting of many CLs, can then be generalized to include some CLs, and some RCLs in the intermediate stages. We call the corresponding network a Convolutional Neural Network with Fully Recursive Perceptron Network (C-FRPN). Through an analysis of results obtained from applications of the C-FRPN to three benchmark image classification datasets: CIFAR-10, SVHN, ISIC, it is found that (i) in general, the performance of a C-FRPN, even with only one RCL, is better than the performance of the corresponding deep CNN with all CLs, under the constraint of having the same number of unknown parameters; (ii) the performance of the C-FRPN varies with respect to (a) where the RCLs are located, and (b) the number of RCLs in the C-FRPN; and, (iii) the effectiveness of the RCLs depends on the size of the training dataset. The results suggest that: (a) it is advisable to use RCLs particularly when training very large sets of data, (b) it is best to prioritize placement of RCLs close to the input layer of the C-FRPN, and (c) it is advisable to increase the number of RCLs as long as the training dataset can sustain without overfitting being observed. [ABSTRACT FROM AUTHOR]

Despite the superiority of convolutional neural networks demonstrated in time series modeling and forecasting, it has not been fully explored on the design of the neural network architecture and the tuning of the hyper-parameters. Inspired by the incremental construction strategy for building a random multilayer perceptron, we propose a novel Error-feedback Stochastic Modeling (ESM) strategy to construct a random Convolutional Neural Network (ESM-CNN) for time series forecasting task, which builds the network architecture adaptively. The ESM strategy suggests that random filters and neurons of the error-feedback fully connected layer are incrementally added to steadily compensate the prediction error during the construction process, and then a filter selection strategy is introduced to enable ESM-CNN to extract the different size of temporal features, providing helpful information at each iterative process for the prediction. The performance of ESM-CNN is justified on its prediction accuracy of one-step-ahead and multi-step-ahead forecasting tasks respectively. Comprehensive experiments on both the synthetic and real-world datasets show that the proposed ESM-CNN not only outperforms the state-of-art random neural networks, but also exhibits stronger predictive power and less computing overhead in comparison to trained state-of-art deep neural network models. [ABSTRACT FROM AUTHOR]

• Developing a new recursive convolutional layer (RCL) able to stop the iteration when the hidden state stabilize. • First network to have different depth for different batch of samples. • A comprehensive study on the best locations and numbers of RCL in a multilayer network called C-FRPN. To overcome problems with the design of large networks, particularly with respect to the depth of the network, this paper presents a new model of convolutional neural networks (CNN) which features fully recursive convolutional layers (RCLs). An RCL is a generalization of the classic one-stage feedforward convolutional layer (CL) to fully direct feedback connections between the outputs of the CL and its inputs. A traditional deep CNN consisting of many CLs, can then be generalized to include some CLs, and some RCLs in the intermediate stages. We call the corresponding network a Convolutional Neural Network with Fully Recursive Perceptron Network (C-FRPN). Through an analysis of results obtained from applications of the C-FRPN to three benchmark image classification datasets: CIFAR-10, SVHN, ISIC, it is found that (i) in general, the performance of a C-FRPN, even with only one RCL, is better than the performance of the corresponding deep CNN with all CLs, under the constraint of having the same number of unknown parameters; (ii) the performance of the C-FRPN varies with respect to (a) where the RCLs are located, and (b) the number of RCLs in the C-FRPN; and, (iii) the effectiveness of the RCLs depends on the size of the training dataset. The results suggest that: (a) it is advisable to use RCLs particularly when training very large sets of data, (b) it is best to prioritize placement of RCLs close to the input layer of the C-FRPN, and (c) it is advisable to increase the number of RCLs as long as the training dataset can sustain without overfitting being observed. [ABSTRACT FROM AUTHOR]

• HELP improves the efficiency of hyperparameter tunning with random exploration. • HELP avoids bad hyperparameters in future exploration by learning with an RNN. • HELP finds the optimal hyperparameters at iteration by information during training. Hyperparameter selection is very important for the success of deep neural network training. Random search of hyperparameters for deep neural networks may take a long time to converge and yield good results because the training of deep neural networks with a huge number of parameters for every selected hyperparameter is very time-consuming. In this work, we propose the Hyperparameter Exploration LSTM-Predictor (HELP) which is an improved random exploring method using a probability-based exploration with an LSTM-based prediction. The HELP has a higher probability to find a better hyperparameter with less time. The HELP uses a series of hyperparameters in a time period as input and predicts the fitness values of these hyperparameters. Then, exploration directions in the hyperparameter space yielding higher fitness values will have higher probabilities to be explored in the next turn. Experimental results for training both the Generative Adversarial Net and the Convolution Neural Network show that the HELP finds hyperparameters yielding better results and converges faster. [ABSTRACT FROM AUTHOR]

Deep neural networks (DNN), such as convolutional neural networks (CNN) have been widely used for object recognition. However, they are usually unable to ensure the required intra-class compactness and inter-class separability in the kernel space. These are known to be important in pattern recognition for achieving both robustness and accuracy. In this paper, we propose to integrate a kernelized Min-Max objective in the DNN training in order to explicitly enforce both kernelized within-class compactness and between-class margin. The involved kernel space is implicitly mapped from the feature space associated with a certain upper layer of DNN by exploiting a kernel trick, while the Min-Max objective in this space is interpolated with the original DNN loss function and finally optimized in the training phase. With a very small additional computation cost, the proposed strategy can be easily integrated in different DNN models without changing any other part of the original model. The comparative recognition accuracy of the proposed method is evaluated with multiple DNN models (including shallow CNN, deep CNN and deep residual neural network models) on two benchmark datasets: CIFAR-10 and CIFAR-100. Extensive experimental results demonstrate that the integration of kernelized Min-Max objective in the training of DNN models can achieve better results compared to state-of-the-art models, without incurring additional model complexity. [ABSTRACT FROM AUTHOR]

Hashing has been widely used for large-scale multimedia retrieval because of its advantages in storage and retrieval efficiency. Traditional supervised hash methods represent an image as a feature vector and then perform a separate quantization step to generate a binary code. Due to the difficulty of discrete optimization of hash codes, continuous relaxation is generally used to replace discrete optimization. However, the process of continuous relaxation leads to inevitable quantization error. To avoid this drawback, a deep balanced discrete hashing method is proposed, which uses discrete gradient propagation with the straight-through estimator. The proposed method does not use the traditional continuous relaxation strategy, thereby reducing the quantization error caused by continuous relaxation. And the proposed method uses supervised information to directly guide the discrete coding and deep feature learning process. In the proposed method, the last layer of the Convolutional Neural Network (CNN) outputs the binary code directly. In the loss function, discrete values are calculated by combining the pairwise loss and a balance controlling term. The learned binary hash code maintains the similar relationship and label consistency at the same time. While maintaining the pairwise similarity, the proposed method keeps the balance of hash codes to improve retrieval performance. Extensive experiments show that the proposed method outperforms the state-of-the-art hashing methods on four image retrieval benchmark datasets. [ABSTRACT FROM AUTHOR]

The deep neural network model based on spatial pyramid pooling for multi-scale context information has been widely adopted in image semantic segmentation. However, this structure reduces the feature resolution during the inference process, which should be recovered by bilinear interpolation. We find that the performance of boundary location heavily depends on the interpolation operation. Assuming that the foreground object has a boundary with uneven curvature changes, it will be difficult to preserve details in the lower resolution feature map. In this case, if the model achieves near-optimal fitting, the prediction of the boundary point needs to satisfy strict constraints. But this condition is very difficult to meet for the model. For this purpose, we propose a constraint balance factor (CBF) relaxation method to improve the fitting ability of the model. The constraint balance factor is learned end-to-end by a sub-network. This sub-network can be easily embedded into a semantic segmentation network that requires interpolation without too many extra parameter increments. The experimental results illustrate that the proposed CBFNet performs well on PASCAL VOC 2012 and Cityscapes datasets. [ABSTRACT FROM AUTHOR]

Visual cortex is able to process information in multiple pathways and integrate various forms of representations. This paper proposed a bio-inspired method that utilizes the line-segment-based representation to perform a dedicated channel for the geometric feature learning process. The extracted geometric information can be integrated with the original pixel-based information and implemented on both the convolutional neural networks (SegCNN) and the stacked autoencoders (SegSAE). Segment-based operations such as segConvolve and segPooling are designed to further process the extracted geometric features. The proposed models are verified on the MNIST dataset, Caltech 101 dataset and QuickDraw dataset for image classification. According to the experimental results, the proposed models can facilitate the classification accuracies especially when the sizes of the training set are limited. Particularly, the method based on multiple representations is found to be effective for classifying the hand-drawn sketches. [ABSTRACT FROM AUTHOR]

Video-based convolutional neural networks (CNNs) involve a large amount of training parameters, leading to the enormous computational complexity, which thereby delays the network convergence. Therefore, training successful CNN for action recognition rapidly is non-trivial. In this paper, a novel encoding called nonlinear gated channels unit (NGCU) is proposed to facilitate network training by encoding the global channel-level relationship. Based on this, nonlinear gated channels network (NGCN) is constructed for the end-to-end encoding, and the corresponding convergence performance is evaluated on the standard benchmarks UCF101 and HMDB51. Experimental results demonstrate that the proposed method is conducive to the convergence process of CNN based action recognition models. [ABSTRACT FROM AUTHOR]

Unlike natural image captured with cameras, screen content image (SCI) is a composite image including textual and pictorial regions. The different characteristics lead to many difficulties in image quality assessment (IQA). Most existing models based on the convolutional neural network (CNN) divide large SCIs into image patches to increase training samples for CNN training. This brings two problems: (1) a single image patch can not represent the quality of the entire image, especially in IQA of SCI; (2) SCI patches of an entire image degraded by the same distortion type and strength may have drastically different quality. In addition, these models adopt the mean square error (MSE) between the predicted quality and the subjective differential mean opinion score (DMOS) to train the CNN, without considering quality ranking between different SCIs. In this paper, we propose a novel no-reference (NR) IQA model based on the convolutional neural network (CNN). The contributions of our algorithm can be concluded as follows: (1) considering a large difference exists in different regions in a SCI, the pseudo global features generated with multi-region local features are utilized for quality evaluation, which better reflect image quality than local features of each image patch; (2) the noise classification task is used as an auxiliary task which aids the quality score prediction task to improve the representation ability; (3) the Siamese networks are used to predict the quality scores of two different SCIs, and a new ranking loss is proposed to rank the predicted scores, aiming to enhance the ability of the model to rank image in terms of quality. Experimental results verify that our model outperforms all test NR IQA methods and full-reference (FR) IQA methods on the screen content image quality assessment database (SIQAD). [ABSTRACT FROM AUTHOR]

Highlights • The re-identification problem is re-formulate as a multi-instance verification problem, to meet the requirement of more realistic real-world applications. • Two paradigms IS and ES are developed to effectively perform multiinstance learning in an end-to-end way, which takes pairs of bags as input and directly output bag labels. • The proposed method relaxes the requirements for strictly accurate annotations, accordingly increases the availability of large re-identification datasets. • The proposed MICNN can achieve superior performance against single instance CNN framework. Abstract This paper tackles the challenging problem of multi-shot person re-identification with Convolutional Neural Network (CNN). As no prior information about how importance each instance plays, it is non-trivial to exploit the interaction information shared by the multi-shot images to help identification. Traditional CNN is in single-shot architecture, then how to utilize the interaction information provided by multi-shot images becomes an important problem to solve. Furthermore, as data augmentation methods are not strictly label-preserving, it increases the difficulty to select discriminative instance for CNN training. In this paper, we propose a weakly supervised CNN framework named Multi-Instance Convolutional Neural Network (MICNN) to solve the aforementioned problem. We develop two paradigms, i.e., Embedding-Space paradigm and Instance-Space paradigm, which re-formulate the person re-identification problem as a multi-instance verification problem with part-based features extracted by neural network. We respectively devise a specific bag-level loss function which incorporates the characteristics of the multi-instance problem for each paradigm. Experiments show that the proposed IS method outperforms many related state-of-the-art techniques on four benchmark datasets: CUHK03, SYSUm, RAiD and Market-1501. [ABSTRACT FROM AUTHOR]