Extract Discriminative Features Research Articles

Performance analysis of deep unified model for facial expression recognition using convolution neural network

Facial expression recognition has gathered substantial attention in computer vision applications, with the need for robust and accurate models that can decipher human emotions from facial images. Performance analysis of a novel hybrid model combines the strengths of residual network (ResNet) and dense network (DenseNet) architectures after applying preprocessing for facial expression recognition. The proposed hybrid model capitalizes on the complementary characteristics of ResNet's and DenseNet's densely connected blocks to enhance the model's capacity to extract discriminative features from facial images. This research evaluates the hybrid model performance and conducts a comprehensive benchmark against established facial expression recognition convolution neural network (CNN) models. The analysis encompasses key aspects of model performance, including classification accuracy, and adaptability with the LFW dataset for facial expressions such as Anger, Fear, Happy, Disgust, Sad, Surprise, along Neutral. The research observes that the proposed hybrid model is more accurate and efficient computationally than existing models consistently. This performance analysis eliminates information on the hybrid model's perspective to further facial expression recognition research.

Hard Sample Mining-Enabled Supervised Contrastive Feature Learning for Wind Turbine Pitch System Fault Diagnosis

Abstract The presence of multiple failure severities in the wind turbine pitch system due to the long-term wear and tear poses challenges in accurately classifying the system's health condition, thus increasing maintenance costs or damage risks. This paper proposes a novel method based on hard sample mining-enabled supervised contrastive learning (HSMSCL) to address this problem. The proposed method leverages the powerful feature extraction capabilities of supervised contrastive learning to extract discriminative features from highly imbalanced data. Additionally, the method incorporates a cosine similarity-based hard sample mining framework that constructs hard sample pairs within mini-batches during both the representation learning and classifier training phases, thereby improving the diagnostic performance of the model for hard samples. The proposed method achieves macro G-mean of 0.9991 and 0.9971 on two real datasets containing data on wind turbine pitch system cog belt fractures. These results indicate significantly superior fault diagnosis performance compared to existing methods, highlighting its potential for enhancing the reliability and efficiency of fault diagnosis in wind turbine pitch systems.

Pest-ConFormer: A hybrid CNN-Transformer architecture for large-scale multi-class crop pest recognition

Crop pests are acknowledged as the major factors in reducing the yield and quality of agricultural production worldwide. It is an urgent necessity to recognize crop pests accurately to protect the crop in the early stage to reduce the loss for the agricultural economy. Due to the ecological characteristics of the crop pests and the complex background in fields, the crop pests show high inter-class similarity and significant intra-class variation in external morphology appearance, which makes current recognition methods suffer from low classification accuracy and poor generalization ability in complex natural environment recognition tasks. To tackle this problem, a hybrid convolutional neural network and transformer-based model, namely Pest-ConFormer, featured with multi-scale weakly supervised feature selection mechanisms is proposed, which has shown excellent multi-scale discriminative feature extraction in fine-grained visual classification (FGVC) tasks. This method employs a hybrid convolution-transformer encoder architecture pre-training in a self-supervised masked autoencoder manner as a backbone to learn pests’ highly discriminative features across various scales. Next, a dual-path feature aggregation structure with a top-down FPN-like feature pathway and a bottom-up PANet-like feature pathway based on attention mechanisms is designed to learn high-level global context information and low-level local detailed feature representation. Thirdly, a fine-grained classification module using weakly supervised learning is introduced to select the discriminative feature points in different pyramidal levels. Then, these feature points are fed into a graph convolutional network to accomplish classification. Several experiments are conducted on the large-scale multi-class IP102 benchmark dataset, and the proposed method achieves an accuracy of 77.81 % regarding crop pest recognition. The experimental results indicate that our approach outperforms other state-of-the-art methods by nearly 2 percent points, demonstrating that the proposed hybrid architecture with dual-path feature aggregation and fine-grained classification modules can be more effective in the crop pest recognition field than CNN-based methods and can be deployed in the practical natural environment. The source code will be available at https://github.com/mwfang/pestconformer.

X-ray pneumonia detection using angular and radial local binary patterns fusion

Detecting pneumonia via X-ray images is a crucial phase in determining any probable bacterial or viral disease. The automation of this stage is critical since it speeds up the diagnosing process. LBPs have excelled in investigating local conditions within a neighborhood, where they aggregate and analyze local statistical results. In this paper, we present a novel family of LBP descriptors based on local accumulative pixel disparities. We construct the new descriptors using a circular neighborhood acquired from the initial square filter. These descriptors consider angular and radial transitions that occur in both the microstructures and macrostructures of image textures. Eventually, a more accurate and comprehensive visual representation is obtained. Furthermore, we offer a data fusion step based on a voting mechanism to integrate the retrieved data efficiently. We show many types of analyses that demonstrate LBP's and the suggested extensions' capability to extract discriminant features from X-ray images. The suggested method is tested on two different datasets with large diversities that include images from various demographics and regions. Several measures, including accuracy rate, precision, sensitivity, F-measure, and specificity, are used to evaluate the system's efficacy. According to the testing results, the proposed system provides a best successful recognition rate of 82.7 %.

Effect of Particle Swarm Optimization Convolutional Neural Network in An Iris Recognition System

An iris recognition system based on Convolutional Neural Network with Particle Swarm Optimization (CNN-PSO) was developed to improve the identified hitches in the existing systems. Iris images of 150 and 108 persons were acquired from LAUIRIS (Nigeria) and CASIA (China) respectively. The images were resized and cropped after which Hough transform was used for effective localization of the iris region and normalised using Daugman’s rubber sheet model, while an efficient Cumulative Sum-based analysis method was used to extract discriminative features from the normalised iris images after which the iris code was generated. The iris code generated in a vector form was optimised with PSO after which they are fed into Convolutional neural network; the same procedure was engaged during enrolment and authentication to generate the iris template. Euclidean distance was used for decision making on test sample template and stored template. The system was implemented with MATLAB R2013a. The performance of the developed system was evaluated on LAURIS and CASIA, and compared with the existing systems (CNN, BPNN-PSO and BPNN) using False Acceptance Rate (FAR), False Rejection Rate (FRR) and Recognition Rate (RR). CNN-PSO has the highest recognition rate of 98.67% and 97.22% for LAUIRIS and CASIA respectively among the systems which showed an improvement over other three recognition technique. The developed CNN-PSO has not only produced an improved Iris recognition system over the others, with the highest recognition rate for both datasets but it also provides a significant recognition rate of black Iris images despite the limitations identified with black Iris images in separating Iris image from other part of the eyes. The developed

Effect of Particle Swarm Optimization Convolutional Neural Network in An Iris Recognition System

An iris recognition system based on Convolutional Neural Network with Particle Swarm Optimization (CNN-PSO) was developed to improve the identified hitches in the existing systems. Iris images of 150 and 108 persons were acquired from LAUIRIS (Nigeria) and CASIA (China) respectively. The images were resized and cropped after which Hough transform was used for effective localization of the iris region and normalised using Daugman’s rubber sheet model, while an efficient Cumulative Sum-based analysis method was used to extract discriminative features from the normalised iris images after which the iris code was generated. The iris code generated in a vector form was optimised with PSO after which they are fed into Convolutional neural network; the same procedure was engaged during enrolment and authentication to generate the iris template. Euclidean distance was used for decision making on test sample template and stored template. The system was implemented with MATLAB R2013a. The performance of the developed system was evaluated on LAURIS and CASIA, and compared with the existing systems (CNN, BPNN-PSO and BPNN) using False Acceptance Rate (FAR), False Rejection Rate (FRR) and Recognition Rate (RR). CNN-PSO has the highest recognition rate of 98.67% and 97.22% for LAUIRIS and CASIA respectively among the systems which showed an improvement over other three recognition technique. The developed CNN-PSO has not only produced an improved Iris recognition system over the others, with the highest recognition rate for both datasets but it also provides a significant recognition rate of black Iris images despite the limitations identified with black Iris images in separating Iris image from other part of the eyes. The developed technique can be applied to various field of life like security, surveillance systems.

Boosting person ReID feature extraction via dynamic convolution

Extraction of discriminative features is crucial in person re-identification (ReID) which aims to match a query image of a person to her/his images, captured by different cameras. The conventional deep feature extraction methods on ReID employ CNNs with static convolutional kernels, where the kernel parameters are optimized during the training and remain constant in the inference. This approach limits the network's ability to model complex contents and decreases performance, particularly when dealing with occlusions or pose changes. In this work, to improve the performance without a significant increase in parameter size, we present a novel approach by utilizing a channel fusion-based dynamic convolution backbone network, which enables the kernels to change adaptively based on the input image, within two existing ReID network architectures. We replace the backbone network of two ReID methods to investigate the effect of dynamic convolution on both simple and complex networks. The first one called Baseline, is a simpler network with fewer layers, while the second, CaceNet represents a more complex architecture with higher performance. Evaluation results demonstrate that both of the designed dynamic networks improve identification accuracy compared to the static counterparts. A significant increase in accuracy is reported under occlusion tested on Occluded-DukeMTMC. Moreover, our approach achieves a performance comparable to the state-of-the-art on Market1501, DukeMTMC-reID, and CUHK03 with a limited computational load. These findings validate the effectiveness of the dynamic convolution in enhancing the person ReID networks and push the boundaries of performance in this domain.

Boosting person ReID feature extraction via dynamic convolution

Extraction of discriminative features is crucial in person re-identification (ReID) which aims to match a query image of a person to her/his images, captured by different cameras. The conventional deep feature extraction methods on ReID employ CNNs with static convolutional kernels, where the kernel parameters are optimized during the training and remain constant in the inference. This approach limits the network's ability to model complex contents and decreases performance, particularly when dealing with occlusions or pose changes. In this work, to improve the performance without a significant increase in parameter size, we present a novel approach by utilizing a channel fusion-based dynamic convolution backbone network, which enables the kernels to change adaptively based on the input image, within two existing ReID network architectures. We replace the backbone network of two ReID methods to investigate the effect of dynamic convolution on both simple and complex networks. The first one called Baseline, is a simpler network with fewer layers, while the second, CaceNet represents a more complex architecture with higher performance. Evaluation results demonstrate that both of the designed dynamic networks improve identification accuracy compared to the static counterparts. A significant increase in accuracy is reported under occlusion tested on Occluded-DukeMTMC. Moreover, our approach achieves a performance comparable to the state-of-the-art on Market1501, DukeMTMC-reID, and CUHK03 with a limited computational load. These findings validate the effectiveness of the dynamic convolution in enhancing the person ReID networks and push the boundaries of performance in this domain.

Detecting Artificial Intelligence-Generated images via deep trace representations and interactive feature fusion

The detection of Artificial Intelligence-Generated (AIG) images plays an important role in verifying the authenticity and originality of digital images. However, recent advancements in state-of-the-art image generation methods have significantly challenged the ability to differentiate AIG images from natural photographs (NP). To address this issue, a novel approach based on deep trace representations and dual-branch interactive feature fusion is presented. Firstly, a global feature extraction module that leverages attention-based MobileViT (AT-MobileViT) is designed to learn the deep representations of the global trace information. Besides, we apply multiple enhanced residual blocks to extract discriminative multi-scale features. After that, a low-level feature extraction module incorporating a channel-spatial attention (CSA) block is also carefully employed to enhance the learning of trace representations. To facilitate the capture of complementary information between features, a dual-branch interactive feature fusion module is introduced by reshaping feature vectors into interactive matrices. By conducting experiments on both seen and unseen images, results demonstrate the better performance and robustness of the proposed method.

RCFNet: Related Cross-level Feature Network with Cascaded Self-distillation for Monocular Depth Estimation

Monocular depth estimation (MDE) is a challenging yet crucial computer vision task, which aims to generate accurate depth maps from a single image. Existing MDE approaches mainly rely on extracting and fusing diverse information from multi-level features to improve prediction accuracy. However, these methods often apply a traditional feature pyramid structure, neglecting a comprehensive exploration of feature fusion paths across multiple levels. Moreover, a single-feature fusion strategy has limited ability to optimize the network. We propose a novel related cross-level feature network (RCFNet) with cascaded self-distillation for monocular depth estimation, including a cross-level feature enhancement (CLFE) module, a hierarchical feature cross refinement (HFCR) module, and a cascaded self-distillation (CSD) module. The CLFE module integrates cross-level features to further exploit the highest-level features, where a channel attention mechanism with hybrid weight operations is deployed to enhance the initial features. The HFCR module adaptively captures strongly correlated complementary information through a window-based multi-head cross-attention mechanism to generate refined features. Meanwhile, the CSD module with a hierarchical feature transformation loss is proposed, which can be viewed as a virtual teacher to progressively extract discriminative features within the network for better gradient flow improvement. Extensive experiments on NYUv2 and KITTI datasets demonstrate that our method outperforms existing SOTA MDE methods in terms of accuracy capacity and robustness.

Multi-scale occlusion suppression network for occluded person re-identification

In practical application scenarios, the occlusion caused by various obstacles greatly undermines the accuracy of person re-identification. Most existing methods for occluded person re-identification focus on inferring visible parts of the body through auxiliary models, resulting in inaccurate feature matching of parts and ignoring the problem of insufficient occluded samples, which seriously affects the accuracy of occluded person re-identification. To address the above issues, we propose a multi-scale occlusion suppression network (MSOSNet) for occluded person re-identification. Specifically, we first propose a dual occlusion augmentation module (DOAM), which combines random occlusion with our proposed novel cross occlusion to generate more diverse occlusion data. Meanwhile, we design a novel occluded-aware spatial attention module (OSAM) to enable the network to focus on non-occluded areas of pedestrian images and effectively extract discriminative features. Ultimately, we propose a part feature matching module (PFMM) that utilizes graph matching algorithms to match non-occluded body parts of pedestrians. Extensive experimental results on both occluded and holistic datasets validate the effectiveness of our method.

GRLN: Gait Refined Lateral Network for gait recognition

Gait recognition aims to identify individuals at a distance based on their biometric gait patterns. While offering flexibility in network input, existing set-based methods often overlook the potential of fine-grained local feature by solely utilizing global gait feature and fail to fully exploit the communication between silhouette-level and set-level features. To alleviate this issue, we propose Gait Refined Lateral Network(GRLN), featuring plug-and-play Adaptive Feature Refinement modules (AFR) that extract discriminative features progressively from silhouette-level and set-level representations in a coarse-to-fine manner at various network depths. AFR can be widely applied in set-based gait recognition models to substantially enhance their gait recognition performance. To align with the extracted refined features, we introduce Horizontal Stable Mapping (HSM), a novel mapping technique that reduces model parameters while improving experimental results. To demonstrate the effectiveness of our method, we evaluate GRLN on two gait datasets, achieving the highest recognition rate among all set-based methods. Specifically, GRLN demonstrates an average improvement of 1.15% over the state-of-the-art set-based method on CASIA-B. Especially in the coat-wearing condition, GRLN exhibits a 5% improvement in performance compared to the contrast method GLN.

Multi-granularity tooth analysis via YOLO-based object detection models for effective tooth detection and classification

Accurate detection and classification of teeth is the first step in dental disease diagnosis. However, the same class of tooth exhibits significant variations in surface appearance. Moreover, the complex geometrical structure poses challenges in learning discriminative features among the different tooth classes. Due to these complex features, tooth classification is one of the challenging research domains in deep learning. To address the aforementioned issues, the presented study proposes discriminative local feature extraction at different granular levels using YOLO models. However, this necessitates a granular intra-oral image dataset. To facilitate this requirement, a dataset at three granular levels (two, four, and seven teeth classes) is developed. YOLOv5, YOLOv6, and YOLOv7 models were trained using 2,790 images. The results indicate superior performance of YOLOv6 for two-class classification problems. The model generated a mean average precision (mAP) value of 94%. However, as the granularity level is increased, the performance of YOLO models decreases. For, four and seven-class classification problems, the highest mAP value of 87% and 79% was achieved by YOLOv5 respectively. The results indicate that different levels of granularity play an important role in tooth detection and classification. The YOLO’s performance gradually decreased as the granularity decreased especially at the finest granular level.

Decoding Innocence: Advancing Forensic Facial Discrimination through Comparative Analysis of Conventional CNN and Advanced Architectures

The field of Digital Image Forensics (DIF) faces a critical issue in accurately identifying children in digital images, notably in cases involving the proliferation of child sexual abuse content. Existing techniques face hurdles due to model architecture limitations, dataset suitability concerns, and classification imbalance, impeding their ability to recognize children to deter pornographic images. Addressing this challenge, this study introduces Implicit Feature Extraction (IFE), a specialized approach for distinguishing child and adult images in object detection. Leveraging Convolutional Neural Networks (CNNs), the IFE method automates the extraction of discriminative facial features, surpassing the constraints of Explicit Feature Extraction (EFE) methods, which achieve an accuracy of around 70%. The research focuses on three core objectives introducing IFE for detailed face feature detection in DIF's child and adult image identification, implementing IFE with CNNs to enhance image classification, and conducting a thorough evaluation of the proposed technique's performance using key metrics like accuracy and balanced classification results and comparing the result with a basic CNN model’s performance. This research's significance lies in its notable contributions to digital image forensics, particularly in combatting child exploitation. The fusion of IFE with CNNs showcases 92% accuracy in distinguishing child and adult images, promising advancements with practical implications in child protection and forensic investigations. The comprehensive evaluation using the UTKFace dataset underscores the proposed technique's efficacy, marking a substantial improvement in child image identification within digital image forensics.

Traffic Sign Classification Using CNN

This study presents a novel approach for traffic sign classification leveraging Convolutional Neural Networks (CNNs). With the proliferation of autonomous vehicles and advanced driver assistance systems, accurate and efficient traffic sign recognition is imperative for safe and efficient navigation. The proposed CNN-based method utilizes deep learning techniques to automatically extract discriminative features from traffic sign images, enabling robust classification across diverse environmental conditions and variations in sign appearance. Experimental results demonstrate the effectiveness of the proposed approach. This research contributes to the advancement of intelligent transportation systems by providing a reliable and scalable solution for real-time traffic sign classification. Our findings highlight the potential of deep learning for robust and accurate traffic sign classification in real-world scenarios.

Multi-culture Sign Language Detection and Recognition Using Fine-tuned Convolutional Neural Network

The speech and hearing-impaired community use sign language as the primary means of communication. It is quite challenging for the general population to interpret or learn sign language completely. A sign language recognition system must be designed and developed to address this communication barrier. Most current sign language recognition systems rely on wearable sensors, keeping the recognition system unaffordable for most individuals. Moreover, the existing vision-based sign recognition frameworks do not consider all of the spatial and temporal information required for accurate recognition. A novel vision-based hybrid deep neural net methodology is proposed in this project for recognizing American Sign Language (ASL) and custom sign gestures. The proposed framework aims to establish a single framework for tracking and extracting multi-semantic properties, such as non-manual components and manual co-articulations. Furthermore, spatial feature extraction from the sign gestures is deployed using a Hybrid Deep Neural Network (HDNN) with atrous convolutions. The temporal and sequential feature extraction is carried out by employing attention-based HDNN. In addition, the distinguished abstract feature extraction is done using modified autoencoders. The discriminative feature extraction for differentiating the sign gestures from unwanted transition gestures is done by leveraging the hybrid attention module. The experimentation of the proposed model has been carried out on the novel multi-signer ASL and custom sign language dataset. The proposed sign language recognition framework with hybrid neural nets, specifically using HDNN, yields better results than other state-of-the-art frameworks. Additionally, a detection module is incorporated using Flask web, allowing manual input of images/signs for real-time recognition.

Employing a Convolutional Neural Network to Classify Sleep Stages from EEG Signals Using Feature Reduction Techniques

One of the most essential components of human life is sleep. One of the first steps in spotting abnormalities connected to sleep is classifying sleep stages. Based on the kind and frequency of signals obtained during a polysomnography test, sleep phases can be separated into groups. Accurate classification of sleep stages from electroencephalogram (EEG) signals plays a crucial role in sleep disorder diagnosis and treatment. This study proposes a novel approach that combines feature selection techniques with convolutional neural networks (CNNs) to enhance the classification performance of sleep stages using EEG signals. Firstly, a comprehensive feature selection process was employed to extract discriminative features from raw EEG data, aiming to reduce dimensionality and enhance the efficiency of subsequent classification using mutual information (MI) and analysis of variance (ANOVA) after splitting the dataset into two sets—the training set (70%) and testing set (30%)—then processing it using the standard scalar method. Subsequently, a 1D-CNN architecture was designed to automatically learn hierarchical representations of the selected features, capturing complex patterns indicative of different sleep stages. The proposed method was evaluated on a publicly available EDF-Sleep dataset, demonstrating superior performance compared to traditional approaches. The results highlight the effectiveness of integrating feature selection with CNNs in improving the accuracy and reliability of sleep stage classification from EEG signals, which reached 99.84% with MI-50. This approach not only contributes to advancing the field of sleep disorder diagnosis, but also holds promise for developing more efficient and robust clinical decision support systems.

DLA-Net: dual lesion attention network for classification of pneumoconiosis using chest X-ray images

Accurate and early detection of pneumoconiosis using chest X-rays (CXR) is important for preventing the progression of this incurable disease. It is also a challenging task due to large variations in appearance, size and location of lesions in the lung regions as well as inter-class similarity and intra-class variance. Compared to traditional methods, Convolutional Neural Networks-based methods have shown improved results; however, these methods are still not applicable in clinical practice due to limited performance. In some cases, limited computing resources make it impractical to develop a model using whole CXR images. To address this problem, the lung fields are divided into six zones, each zone is classified separately and the zone classification results are then aggregated into an image classification score, based on state-of-the-art. In this study, we propose a dual lesion attention network (DLA-Net) for the classification of pneumoconiosis that can extract features from affected regions in a lung. This network consists of two main components: feature extraction and feature refinement. Feature extraction uses the pre-trained Xception model as the backbone to extract semantic information. To emphasise the lesion regions and improve the feature representation capability, the feature refinement component uses a DLA module that consists of two sub modules: channel attention (CA) and spatial attention (SA). The CA module focuses on the most important channels in the feature maps extracted by the backbone model, and the SA module highlights the spatial details of the affected regions. Thus, both attention modules combine to extract discriminative and rich contextual features to improve classification performance on pneumoconiosis. Experimental results show that the proposed DLA-Net outperforms state-of-the-art methods for pneumoconiosis classification.

Spectral Spatial Neighborhood Attention Transformer for Hyperspectral Image Classification

Hyperspectral image (HSI) can provide rich spectral information, which can be helpful for accurate classification in many applications. However, the hyperspectral image classification task has challenges, including limited labeled data, data redundancy, data sparsity, and imbalanced class samples. Over time, various methods have been proposed to solve the above-mentioned problems. To mitigate the issue mentioned above in this paper, we present the neighborhood attention transformer with a channel-wise shift technique for hyperspectral image classification. The neighborhood attention transformer leverages the power of the attention mechanism to capture the spatial relationship between neighboring pixels and extract discriminative features. The channel-wise shift techniques empower the model to adjust the spectral characteristics of each channel adaptively, enhancing its ability to handle the spectral variations present in hyperspectral data. To validate the effectiveness of the proposed model, we conduct comprehensive experiments on the publically available dataset. The results demonstrate that our model consistently outperforms other state-of-the-art methods. The overall accuracy of the proposed model reached 99.41 % , 99.93 % , 98.35 % 99.59 % four datasets.

Handwritten Character Recognition System

Digitizing handwritten documents and enabling efficient information processing and retrieval require systems that can recognize handwritten characters. This research offers a unique approach for handwritten character detection using state-of-the-art machine learning algorithms. The proposed technique automatically extracts discriminative features from photos of handwritten characters using convolutional neural networks (CNNs). These attributes are then used by a classifier to determine which characters are related. The dataset used for training and assessment is made up of a large collection of handwritten characters gathered under various writing styles, sizes, and orientations in order to guarantee the durability and generalization power of the model. To enhance its quality and diversity, the training data is put through a rigorous preparation procedure that includes picture augmentation, noise removal, and normalization. The studies' results demonstrate how well and precisely the proposed system can recognize handwritten characters in a range of languages and writing styles. The system performs competitively compared to state-of-the-art methods and demonstrates robustness against variations in handwriting style and quality. Furthermore, the system has potential in terms of efficiency and scalability, making it suitable for real-time applications such as document digitalization, handwritten word recognition in electronic devices, and automatic form processing.