Search

No-reference (NR) image quality assessment (IQA) is an important task of computer vision. Most NR-IQA methods via deep neural networks do not reach desirable IQA performance and have bulky models which make them difficult to be used in the practical scenarios. This paper proposes a lightweight transformer and multi-head prediction network for NR-IQA. The proposed method consists of two lightweight modules: feature extraction and multi-head prediction. The module of feature extraction exploits lightweight transformer blocks to learn features at different scales for measuring different image distortions. The module of multi-head prediction uses three weighted prediction blocks and an FC layer to aggregate the learned features for predicting image quality score. The weighted prediction block can measure the importance of different elements of input feature at the same scale. Since the importance of feature elements at the same scale and the importance of the features at different scales are both considered, the module of multi-head prediction can provide more accurate prediction results. Extensive experiments on the standard IQA datasets are conducted. The results show that the proposed method outperforms some baseline NR-IQA methods in IQA performance on the large image datasets. For the model complexity, the proposed method is also superior to several recent NR-IQA methods. [ABSTRACT FROM AUTHOR]

In the previous works on heat stress detection, various state-of-the-art machine learning techniques had been used to detect stress patterns from electroencephalographic (EEG) signals. Since the handcrafted feature engineering-based approaches pose certain limitations and are sensitive to transform the nonlinearities of EEG, deep learning techniques have drawn attention in the domain of EEG-based applications. Moreover, existing approaches for stress detection consider the whole frequency band (delta to gamma) which conceals the redundant and lossy information that increased the false detection rate. Also, these approaches were implemented only to identify the stress in binary classes and confined their application to identifying the level of stress. Therefore, a multi-scale channel-wise convolution-based multi-level heat stress assessment has been proposed in this paper. The novel contributions of this work are (1) designing a multi-scale convolutional neural network (MS-CNN) for extracting precise information from individual frequency bands of EEG, (2) use of sparse connectivity to reduce the redundant information and increase the performance with less trainable parameters, and (3) multi-level heat stress assessment for precise identification of the severity of stress. The proposed approaches were evaluated on pre-recorded data of 10 rats in a simulated laboratory environment. The high accuracy of approximately 96–97% and 90–92% has been achieved for binary and multi-level stress detection. There is approximately a 6 to 50% reduction in the trainable parameters with a 2% improvement of accuracy achieved on adopting channel-wise convolution over MS-CNN and deep convolutional neural network (DCNN). This demonstrates the effectiveness of the adopted multiscale feature extraction and sparse connectivity to improve the performance with a less complex model. [ABSTRACT FROM AUTHOR]

Diabetic Retinopathy (DR), one of the most progressive sight-threatening diseases caused by the long-term diabetic condition, can lead to vision impairment and blindness later. Early diagnosis and timely treatment help control and avert DR from its progression. However, manual grading is exceptionally challenging and arduous due to the complex anatomical features in the retina. Therefore, developing an automated diagnostic method for screening DR is obligatory. This paper proposes a deep learning-enabled optimized feature selection approach to classify the stage of DR severity in the fundus image. At first, a pre-processing phase eradicates the noise and improves the contrast in the retinal fundus image. Then, blood vessel segmentation is performed using the Coherence Enhancing Energy Based Regularized Level Set Evolution method in the green channel fundus image. Subsequently, the optic disk is segmented with Canny Anisotropic Diffusion filter and morphological transformations. Next, the candidate lesion region is detected using an Attention-based Fusion Network (AFU-Net). Then, shape and texture features are extracted, and then, the optimal subset of features is selected using the Improved Harris Hawk Optimization algorithm. Finally, a deep Convolutional Neural Network classifies the DR stages, and the model weight is updated using the same algorithm. The proposed method achieved superior performance in two benchmark public datasets compared with the existing state-of-the-art methods using F1-score, accuracy, sensitivity, and specificity measures. [ABSTRACT FROM AUTHOR]

The convolutional neural network has been proven to be a robust recognition and diagnosis model for modelling diseases. In general, a convolutional neural network consists of feature learning which is considered a vital process in terms of feature extraction and classifier algorithms used for decision-making. In recent years, various pooling operations considered the spine of the feature learning layer have been developed in the literature; unfortunately, few of them can be spread to tackle the overfitting problem. In this paper, a novel pooling approach termed whole image average pooling is proposed to hold over the most significant feature information on the learning layer. The proposed method consists of two main stages: in the first stage, the average of all the values of the input image is calculated, in the second stage, the input image is divided into regions that represent the feature set of the input image, and then, the values of the regions are calculated. While obtaining the values of these features, the average of all values of the input image is taken into account. The proposed whole image average pooling approach addresses the overfitting problem by processing all pixel values in the image data and is used to get the important feature information from the medical datasets. To evaluate the performance of the proposed whole image average pooling, four clinically applicable tumour datasets as medical datasets have been trained and tested on seven convolutional neural network models AlexNet, VGG16, ResNet18, GoogleNet, ResNet50, R-CNN, and Faster R-CNN. The experimental outcomes demonstrate that the proposed whole image average pooling consistently outperformed the state-of-the-art pooling operation in terms of accuracy, sensitivity, specificity, positive predictive value, false-positive rate, and F1-measure evaluation criteria in three datasets. [ABSTRACT FROM AUTHOR]

In recent years, the Internet of Things (IoT) has received a lot of attention. It has been used in many applications such as the control industry, industrial plants, and medicine. In this regard, a fundamental necessity is to implement security in IoT. To this end, Network intrusion detection systems (NIDSs) have been recently in the detection of network attacks and threats. Currently, these systems use a variety of deep learning (DL) models such as the convolutional neural networks to improve the detection of attacks. However, almost all current DL-based NIDSs are made up of many layers, and therefore, they need a lot of processing resources because of their high number of parameters. On the other hand, due to the lack of processing resources, such inefficient DL models are unusable in IoT devices. This paper presents a very accurate NIDS that is named DFE, and it uses a very lightweight and efficient neural network based on the idea of deep feature extraction. In this model, the input vector of the network is permuted in a 3D space, and its individual values are brought close together. This allows the model to extract highly discriminative features using a small number of layers without the need to use large 2D or 3D convolution filters. As a result, the network can achieve an accurate classification using a significantly small number of required calculations. This makes the DFE ideal for real-time intrusion detection by IoT devices with limited processing capabilities. The efficacy of the DFE has been evaluated using three popular public datasets named UNSW-NB15, CICIDS2017, and KDDCup99, and the results show the superiority of the proposed model over the state-of-the-art algorithms [ABSTRACT FROM AUTHOR]

In this paper, we investigate Hebbian learning strategies applied to Convolutional Neural Network (CNN) training. We consider two unsupervised learning approaches, Hebbian Winner-Takes-All (HWTA), and Hebbian Principal Component Analysis (HPCA). The Hebbian learning rules are used to train the layers of a CNN in order to extract features that are then used for classification, without requiring backpropagation (backprop). Experimental comparisons are made with state-of-the-art unsupervised (but backprop-based) Variational Auto-Encoder (VAE) training. For completeness,we consider two supervised Hebbian learning variants (Supervised Hebbian Classifiers—SHC, and Contrastive Hebbian Learning—CHL), for training the final classification layer, which are compared to Stochastic Gradient Descent training. We also investigate hybrid learning methodologies, where some network layers are trained following the Hebbian approach, and others are trained by backprop. We tested our approaches on MNIST, CIFAR10, and CIFAR100 datasets. Our results suggest that Hebbian learning is generally suitable for training early feature extraction layers, or to retrain higher network layers in fewer training epochs than backprop. Moreover, our experiments show that Hebbian learning outperforms VAE training, with HPCA performing generally better than HWTA. [ABSTRACT FROM AUTHOR]

Although convolutional neural networks have achieved success in the field of image classification, there are still challenges in the field of agricultural product quality sorting such as machine vision-based jujube defects detection. The performance of jujube defect detection mainly depends on the feature extraction and the classifier used. Due to the diversity of the jujube materials and the variability of the testing environment, the traditional method of manually extracting the features often fails to meet the requirements of practical application. In this paper, a jujube sorting model in small data sets based on convolutional neural network and transfer learning is proposed to meet the actual demand of jujube defects detection. Firstly, the original images collected from the actual jujube sorting production line were pre-processed, and the data were augmented to establish a data set of five categories of jujube defects. The original CNN model is then improved by embedding the SE module and using the triplet loss function and the center loss function to replace the softmax loss function. Finally, the depth pre-training model on the ImageNet image data set was used to conduct training on the jujube defects data set, so that the parameters of the pre-training model could fit the parameter distribution of the jujube defects image, and the parameter distribution was transferred to the jujube defects data set to complete the transfer of the model and realize the detection and classification of the jujube defects. The classification results are visualized by heatmap through the analysis of classification accuracy and confusion matrix compared with the comparison models. The experimental results show that the SE-ResNet50-CL model optimizes the fine-grained classification problem of jujube defect recognition, and the test accuracy reaches 94.15%. The model has good stability and high recognition accuracy in complex environments. [ABSTRACT FROM AUTHOR]

Recent advancements in multielectrode methods and spike-sorting algorithms enable the in vivo recording of the activities of many neurons at a high temporal resolution. These datasets offer new opportunities in the investigation of the biological neural code, including the direct testing of specific coding hypotheses, but they also reveal the limitations of present decoder algorithms. Classical methods rely on a manual feature extraction step, resulting in a feature vector, like the firing rates of an ensemble of neurons. In this paper, we present a recurrent neural-network-based decoder and evaluate its performance on experimental and artificial datasets. The experimental datasets were obtained by recording the auditory cortical responses of rats exposed to sound stimuli, while the artificial datasets represent preset encoding schemes. The task of the decoder was to classify the action potential timeseries according to the corresponding sound stimuli. It is illustrated that, depending on the coding scheme, the performance of the recurrent-network-based decoder can exceed the performance of the classical methods. We also show how randomized copies of the training datasets can be used to reveal the role of candidate spike-train features. We conclude that artificial neural network decoders can be a useful alternative to classical population vector-based techniques in studies of the biological neural code. [ABSTRACT FROM AUTHOR]

In this study, efficient deep transfer learning models are proposed to classify six types of volcanic rocks, and this paper has a novelty in classifying volcanic rock types for the first time using thin section images. Convolutional neural network-based DenseNet121 and ResNet50 networks, which are transfer learning methods, are used to extract the features from thin section images of rocks, and the classification process is carried out with a single-layer fully connected neural network. The proposed models are trained and tested on 1200 thin section images using four different optimizers (Adadelta, ADAM, RMSprop, SGD). AUC, accuracy, precision, recall and f1-score are used as performance metrics. Proposed models are run 10 times for each optimizer. DenseNet121 classifies volcanic rock types using RMSprop with an average accuracy of 99.50% and a maximum of 100.00%, and ResNet50 classifies using ADAM with an average accuracy of 98.80% and a maximum of 99.72%. Thus, the applied deep transfer learning is promising in geosciences and can be used to identify rock types quickly and accurately. [ABSTRACT FROM AUTHOR]

Wearable technology offers a prospective solution to the increasing demand for activity monitoring in pervasive healthcare. Feature extraction and selection are crucial steps in activity recognition since it determines the accuracy of activity classification. However, existing feature extraction and selection methods involve manual feature engineering, which is time-consuming, laborious and prone to error. Therefore, this paper proposes an unsupervised feature learning method that automatically extracts and selects the features without human intervention. Specifically, the proposed method jointly trains a convolutional denoising autoencoder with a convolutional neural network to learn the underlying features and produces a compact feature representation of the data. This allows not only more accurate and discriminative features to be extracted but also reduces the computational cost and improves generalization of the classification models. The proposed method was evaluated and compared with deep learning convolutional neural networks on a public dataset. Results have shown that the proposed method can learn a salient feature representation and subsequently recognize the activities with an accuracy of 0.934 and perform comparably well to the convolutional neural networks. [ABSTRACT FROM AUTHOR]

A computerized coconut tree detection system can help dendrologists and laypersons in identifying coconut trees based on three morphological parameters including height, inclination, and orientation. These three parameters help to determine the health and the nature of growth of coconut trees which influences the design and use of robots for harvesting coconuts. Deep learning is a powerful tool used for feature extraction as it is better in extracting deeper details (features) in an image. In this research work, a new Modified Inception Net based Hyper Tuning Support Vector Machine classification method named MIN-SVM is proposed for coconut tree classification based on three morphological parameters including height, inclination and orientation. The features from the pre-processed coconut tree images were extracted using four distinct Convolutional Neural Network models including Visual Geometry Group, Inception Net, ResNet, and MIN-SVM. These extracted features were then classified using a Machine Learning model named Support Vector Machine (SVM). The MIN-SVM have achieved a remarkable accuracy of 95.35 percent as contrasted to Visual Geometry Group (91.90%), Inception Net (81.66%), and ResNet (71.95%). The features extracted from Modified Inception Net fitted good with SVM classifier. Experimental results show that MIN-SVM can be powerful computerized automated system to identify coconut trees based on height, inclination, and orientation. [ABSTRACT FROM AUTHOR]

The process of object tracking involves consistently identifying each instance across frames depending on initial set of object detection(s). Moreover, in multiple object tracking (MOT), the process through tracking-by-detection paradigm consists of performing two common steps consecutively, which are detection and data association. In MOT, it is targeted to associate detections across frames by localizing and identifying all objects of interest. MOT algorithms further keep tracking even the most challenging issues such as revisiting the same view, missing detections, occlusion and temporarily unseen objects, same-appearance objects coexisting in the same frame occur. Hence, re-identification (re-id) appears to be the most powerful tool for assigning the correct identities to each individual instance when aforementioned issues arise. In this work, we propose a similarity-based person re-id framework, called SAT, using a Siamese neural network via shared weights. Once detections are obtained from the backbone SAT applies a Siamese feature extraction model and then we introduce a similarity array for assessing tracklet(s) and detection(s). We examine the performance of SAT on several benchmarks with extensive experiments and statistical tests, where we improve the current state-of-the-art according to commonly used performance metrics with higher accuracy, less ID switches, less false positive and negative rates. [ABSTRACT FROM AUTHOR]

Artificial neural network training algorithms aim to optimize the network parameters regarding the pre-defined cost function. Gradient-based artificial neural network training algorithms support iterative learning and have gained immense popularity for training different artificial neural networks end-to-end. However, training through gradient methods is time-consuming. Another family of training algorithms is based on the Moore–Penrose inverse, which is much faster than many other gradient methods. Nevertheless, most of those algorithms are non-iterative and thus do not support mini-batch learning in nature. This work extends two non-iterative Moore–Penrose inverse-based training algorithms to enable online sequential learning: a single-hidden-layer autoencoder training algorithm and a sub-network-based classifier training algorithm. We further present an approach that uses the proposed autoencoder for self-supervised dimension reduction and then uses the proposed classifier for supervised classification. The experimental results show that the proposed approach achieves satisfactory classification accuracy on many benchmark datasets with extremely low time consumption (up to 50 times faster than the support vector machine on CIFAR 10 dataset). [ABSTRACT FROM AUTHOR]

In the sign language alphabet, several hand signs are in use. Automatic recognition of performed hand signs can facilitate the communication between hearing and none hearing people. This framework proposes hand posture recognition of the American Sign Language alphabet based on a neural network (NN) which works on geometrical feature extraction of the hand. The user's hand is captured by a 3D depth-based sensor camera. Consequently, the hand is segmented according to the depth features. The proposed system is called 'Depth-based Geometrical Sign Language Recognition' (DGSLR). The DGSLR adopted an easier hand segmentation approach, which is further used in other segmentation applications. The proposed geometrical feature extraction framework improves the accuracy of recognition due to unchangeable features against hand orientation or rotation compared to Discrete Cosine Transform (DCT) and Moment Invariant. As a support vector machine (SVM) is a type of artificial neural network (ANN), it is used to drive desired outcomes. Since there are 26 different signs in the Sign Language alphabet, a multi-class SVM versus a single SVM classifier with 26 classes by an RBF kernel was used to validate each class. The proposed framework is proficient to hand posture recognition and provides an accuracy of up to 96.78%. The findings of the iterations demonstrated that the combination of the extracted features resulted in a better accuracy rate in the recognition process in the classification step. [ABSTRACT FROM AUTHOR]

Deep learning has been successfully applied in process monitoring in recent years due to its powerful feature extraction. However, these monitoring methods are difficult to extract intrinsic representations of the process data in complex nonlinear processes. A new deep neural network, residual attention convolutional autoencoder (RACAE) is proposed for process monitoring. The unsupervised learning method of RACAE can extract representative features from high-dimensional data, which can significantly improve process monitoring performance in nonlinear processes. RACAE effectively integrates convolution calculation with an autoencoder to perform effective feature extraction of multivariate data. Moreover, residual attention block is embedded in the autoencoder to select these key features and then reduce the feature dimension for detector. A new process monitoring model is proposed and two kinds of statistics are developed for fault detection. The effectiveness of RACAE in fault detection is evaluated through a numerical case and two benchmark processes. The convolutional autoencoder based on residual attention provides a new approach for feature learning and process monitoring of complex processes. [ABSTRACT FROM AUTHOR]