Robust Deep Learning Research Articles

Overcome the Fear Of Missing Out: Active sensing UAV scanning for precision agriculture

This paper deals with the problem of informative path planning for a UAV deployed for precision agriculture applications. First, we observe that the “fear of missing out” data lead to uniform, conservative scanning policies over the whole agricultural field. Consequently, employing a non-uniform scanning approach can mitigate the expenditure of time in areas with minimal or negligible real value, while ensuring heightened precision in information-dense regions. Turning to the available informative path planning methodologies, we discern that certain methods entail intensive computational requirements, while others necessitate training on an ideal world simulator. To address the aforementioned issues, we propose an active sensing coverage path planning approach, named OverFOMO, that regulates the speed of the UAV in accordance with both the relative quantity of the identified classes, i.e. crops and weeds, and the confidence level of such detections. To identify these instances, a robust Deep Learning segmentation model is deployed. The computational needs of the proposed algorithm are independent of the size of the agricultural field, rendering its applicability on modern UAVs quite straightforward. The proposed algorithm was evaluated with a simu-realistic pipeline, combining data from real UAV missions and the high-fidelity dynamics of AirSim simulator, showcasing its performance improvements over the established state of affairs for this type of missions. An open-source implementation of the algorithm and the evaluation pipeline is also available: https://github.com/emmarapt/OverFOMO.

Analysis and comparison of various deep learning models to implement suspicious activity recognition in CCTV surveillance

The paper aims to analyze and compare various deep learning (DL) algorithms in order to develop a Suspicious Activity Recognition (SAR) system for closed-circuit television (CCTV) surveillance. Automated systems for detecting and classifying suspicious activities are crucial as technology’s role in safety and security expands. This paper addresses these challenges by creating a robust SAR system using machine learning techniques. It analyzes and compares evaluation metrics such as Precision, Recall, F1 Score, and Accuracy using various deep learning methods (convolutional neural network (CNN), Long short-term memory (LSTM) – Visual Geometry Group 16 (VGG16), LSTM – ResNet50, LSTM – EfficientNetB0, LSTM – InceptionNetV3, LSTM – DenseNet121, and Long-term Recurrent Convolutional Network (LRCN)). The proposed system improves threat identification, vandalism deterrence, fight prevention, and video surveillance. It aids emergency response by accurately classifying suspicious activities from CCTV footage, reducing reliance on human security personnel and addressing limitations in manual monitoring. The objectives of the paper include analyzing existing works, extracting features from CCTV videos, training robust deep learning models, evaluating algorithms, and improving accuracy. The conclusion highlights the superior performance of the LSTM-DenseNet121 algorithm, achieving an overall accuracy of 91.17% in detecting suspicious activities. This enhances security monitoring capabilities and reduces response time. Limitations of the system include subjectivity, contextual understanding, occlusion, false alarms, and privacy concerns. Future improvements involve real-time object tracking, collaboration with law enforcement agencies, and performance optimization. Ongoing research is necessary to overcome limitations and enhance the effectiveness of CCTV surveillance.

Invariance embedded physics-infused deep neural network-based sub-grid scale models for turbulent flows

In this paper, we present two novel physics-infused neural network (NN) architectures that satisfy various invariance conditions for constructing efficient and robust Deep Learning (DL)-based sub-grid scale (SGS) turbulence models for use in Large Eddy Simulation (LES) procedures widely used in various fluid engineering applications. The first architecture, called tensor basis neural networks (TBNN), recently proposed in the context of Reynolds-averaged Navier–Stokes (RANS) modeling, and introduced herein for SGS modeling of wall-bounded turbulence, embeds the analytical expansion of the SGS turbulence stresses into integrity basis tensors composed of the symmetric and anti-symmetric parts of the resolved velocity gradient tensor, thus incorporating the Galilean, rotational and reflectional invariances. In our second approach, a relatively simple yet powerful architecture, called Galilean Invariance embedded Neural Network (GINN) incorporates the Galilean invariance, and takes in as inputs the independent components of the integrity basis tensors in addition to the invariant inputs in a single input layer. Explicit filtering of data from direct simulations of the canonical channel flow at two friction Reynolds numbers Reτ≈395 and 590 provided accurate data for training and testing these models. Both sets of models are used to predict the SGS stresses for feature datasets generated with different filter widths, and at different Reynolds numbers. The GINN model yields less prediction error on test datasets (mean squared error ∼0.4) compared to the TBNN model (mean squared error ∼0.5). Upon comparison, it is revealed that GINN has better feature extraction capacity owing to its ability to establish relations between and extract information from cross-components of the integrity basis tensors and the SGS stresses. Based on their predictive performance, both models proposed herein have shown great promise for use in a posteriori actual LESs. The present work illustrates the significance of physics infusion as well as invariance embedding into NN architecture for constructing efficient and robust DL-based SGS turbulent models in achieving superior predictive efficacy.

Lifelong learning with deep conditional generative replay for dynamic and adaptive modeling towards net zero emissions target in building energy system

Deep learning has been advocated as the predominant modeling method in the next-generation green building energy systems for energy prediction, predictive maintenance and control optimization. However, in response to external changes, the limited adaptability to new contents and catastrophic forgetting of previously learnt knowledge result in diminished accuracy and robustness, significantly blocking its practical application. To this end, a novel lifelong learning method with deep generative replay was proposed for dynamic and adaptive modeling to conserve energy and mitigate emissions in building energy systems. The presented lifelong learning method was characterized by the alternate training of the task solver and the replay generator in sequential energy task learning to alleviate the catastrophic forgetting. The replay generator provided the past data for the task solver to retain previous energy knowledge while learn new information, which was a conditional generative model instead of explicitly storing data to save resources and protect privacy. In order to validate its technical feasibility, a field experiment was conducted in a specially constructed net zero energy building for the case study on solar power generation prediction. The overall accuracy of proposed method was 53.4% higher than the standard method through fine-tuning and reached 0.89, which closely approaches the theoretical upper bound of 0.91 obtained by the joint training. Moreover, the proposed method effectively retained previously learnt knowledge in sequential energy task learning, evidenced by an average forgetting rate lower than 0.10. Furthermore, extensive comparative experiments have demonstrated the superiority of the proposed method over other machine learning models based on retraining or incremental training. Our study is expected to develop more flexible and robust deep learning models for improving energy efficiency and promoting the carbon neutrality in building energy systems.

A Tale of Two Cities: Data and Configuration Variances in Robust Deep Learning

A Tale of Two Cities: Data and Configuration Variances in Robust Deep Learning

Multi-modal analysis of infant cry types characterization: Acoustics, body language and brain signals

BackgroundInfant crying is the first attempt babies use to communicate during their initial months of life. A misunderstanding of the cry message can compromise infant care and future neurodevelopmental process. MethodsAn exploratory study collecting multimodal data (i.e., crying, electroencephalography (EEG), near-infrared spectroscopy (NIRS), facial expressions, and body movements) from 38 healthy full-term newborns was conducted. Cry types were defined based on different conditions (i.e., hunger, sleepiness, fussiness, need to burp, and distress). Statistical analysis, Machine Learning (ML), and Deep Learning (DL) techniques were used to identify relevant features for cry type classification and to evaluate a robust DL algorithm named Acoustic MultiStage Interpreter (AMSI). ResultsSignificant differences were found across cry types based on acoustics, EEG, NIRS, facial expressions, and body movements. Acoustics and body language were identified as the most relevant ML features to support the cause of crying. The DL AMSI algorithm achieved an accuracy rate of 92%. ConclusionsThis study set a precedent for cry analysis research by highlighting the complexity of newborn cry expression and strengthening the potential use of infant cry analysis as an objective, reliable, accessible, and non-invasive tool for cry interpretation, improving the infant-parent relationship and ensuring family well-being.

Systemization of Knowledge: Robust Deep Learning using Hardware-software co-design in Centralized and Federated Settings

Deep learning (DL) models are enabling a significant paradigm shift in a diverse range of fields, including natural language processing and computer vision, as well as the design and automation of complex integrated circuits. While the deep models – and optimizations based on them, e.g., Deep Reinforcement Learning (RL) – demonstrate a superior performance and a great capability for automated representation learning, earlier works have revealed the vulnerability of DL to various attacks. The vulnerabilities include adversarial samples, model poisoning, and fault injection attacks. On the one hand, these security threats could divert the behavior of the DL model and lead to incorrect decisions in critical tasks. On the other hand, the susceptibility of DL to potential attacks might thwart trustworthy technology transfer as well as reliable DL deployment. In this work, we investigate the existing defense techniques to protect DL against the above-mentioned security threats. Particularly, we review end-to-end defense schemes for robust deep learning in both centralized and federated learning settings. Our comprehensive taxonomy and horizontal comparisons reveal an important fact that defense strategies developed using DL/software/hardware co-design outperform the DL/software-only counterparts and show how they can achieve very efficient and latency-optimized defenses for real-world applications. We believe our systemization of knowledge sheds light on the promising performance of hardware-software co-design of DL security methodologies and can guide the development of future defenses.

A Robust Deep Learning Approach for Accurate Segmentation of Cytoplasm and Nucleus in Noisy Pap Smear Images

Cervical cancer poses a significant global health burden, affecting women worldwide. Timely and accurate detection is crucial for effective treatment and improved patient outcomes. The Pap smear test has long been a standard cytology screening method, enabling early cancer diagnosis. However, to enhance quantitative analysis and refine diagnostic capabilities, precise segmentation of the cervical cytoplasm and nucleus using deep learning techniques holds immense promise. This research focuses on addressing the primary challenge of achieving accurate segmentation in the presence of noisy data commonly encountered in Pap smear images. Poisson noise, a prevalent type of noise, corrupts these images, impairing the precise delineation of the cytoplasm and nucleus. Consequently, segmentation boundaries become indistinct, leading to compromised overall accuracy. To overcome these limitations, the utilization of U-Net, a deep learning architecture specifically designed for automatic segmentation, has been proposed. This approach aims to mitigate the adverse effects of Poisson noise on the digitized Pap smear slides. The evaluation of the proposed methodology involved a dataset of 110 Pap smear slides. The experimental results demonstrate that the proposed approach successfully achieves precise segmentation of the nucleus and cytoplasm in noise-free images. By preserving the boundaries of both cellular components, the method facilitates accurate feature extraction, thus contributing to improved diagnostic capabilities. Comparative analysis between noisy and noise-free images reveals the superiority of the presented approach in terms of segmentation accuracy, as measured by various metrics, including the Dice coefficient, specificity, sensitivity, and intersection over union (IoU). The findings of this study underline the potential of deep-learning-based segmentation techniques to enhance cervical cancer diagnosis and pave the way for improved quantitative analysis in this critical field of women’s health.

Feature extraction of multi-sensors for early bearing fault diagnosis using deep learning based on minimum unscented kalman filter

Bearing fault diagnosis is vital for ensuring reliability and safety of high-speed trains and wind turbines. Therefore, a minimum unscented Kalman filter-aided deep belief network is proposed to extract invariant features from vibration signals collected by multiple sensors. This particularly crucial due to the dynamic nature of environmental noise and internal bearing degradation, which pose challenges to accurate diagnosis. Firstly, the Gramian angular summation field is employed to transform the multi-sensor signals into 2-D feature maps. This transformation retains the absolute temporal relation within the time-series signals, mitigating feature distortion and enhancing noise elimination for early detection. Secondly, a deep belief network is utilized to construct a robust deep learning framework capable of analysing the translated 2-D feature maps for effective diagnosis. In addition, a minimum unscented transform technique and an adaptive scaling process for noise are integrated into the diagnostic model. These components exhibit exceptional dynamic tracking capabilities, allowing for adjustment of key parameters in response to the prolonged and evolving bearing degradation process. The proposed methodology was rigorously evaluated through a comprehensive analysis involving nine distinct methods, utilizing two diverse bearing datasets. The results obtained underscore the superior attributes. Notably, the proposed diagnostic scheme achieved accuracy rates exceeding 98% and 99% for the two datasets, respectively. This achievement underscores the establishment of an intelligent diagnosis model characterized by high precision and exceptional generalisation capabilities for bearings within rotating machinery. Consequently, this work lays a robust foundation for future research endeavours, particularly in the realm of transfer learning.

CMW-Net: Learning a Class-Aware Sample Weighting Mapping for Robust Deep Learning.

Modern deep neural networks can easily overfit to biased training data containing corrupted labels or class imbalance. Sample re-weighting methods are popularly used to alleviate this data bias issue. Most current methods, however, require to manually pre-specify the weighting schemes relying on the characteristics of the investigated problem and training data. This makes them fairly hard to be generally applied in practical scenarios, due to their significant complexities and inter-class variations of data bias. To address this issue, we propose a meta-model capable of adaptively learning an explicit weighting scheme directly from data. Specifically, by seeing each training class as a separate learning task, our method aims to extract an explicit weighting function with sample loss and task/class feature as input, and sample weight as output, expecting to impose adaptively varying weighting schemes to different sample classes based on their own intrinsic bias characteristics. Extensive experiments substantiate the capability of our method on achieving proper weighting schemes in various data bias cases, like class imbalance, feature-independent and dependent label noises, and more complicated bias scenarios beyond conventional cases. Besides, the task-transferability of the learned weighting scheme is also substantiated, by readily deploying the weighting function learned on relatively smaller-scale CIFAR-10 dataset on much larger-scale full WebVision dataset. The general availability of our method for multiple robust deep learning issues, including partial-label learning, semi-supervised learning and selective classification, has also been validated. Code for reproducing our experiments is available at https://github.com/xjtushujun/CMW-Net.

Energy Harvesting Reconfigurable Intelligent Surface for UAV Based on Robust Deep Reinforcement Learning

Integrating unmanned aerial vehicles with RIS (UAV–RIS) can offer ubiquitous deployment services in communication-disabled areas, but is limited by the on-board energy of the UAVs. In this paper, a novel energy harvesting (EH) scheme on top of the UAV–RIS system, called EH-RIS scheme, is developed for the next generation high performance wireless system. The proposed EH-RIS scheme extends the simultaneous wireless information and power transfer (SWIPT) system by splitting the passive reflected arrays on the geometric space for transporting information and harvesting energy simultaneously. However, pedestrian mobility, and rapid channel changes post challenges to efficient resource allocation in wireless systems. Thus, a robust deep reinforcement learning (DRL)-based algorithm is developed to improve the proposed EH-RIS scheme for guaranteeing the quality of service (QoS) under dynamic wireless environments. The simulation results demonstrate the effectiveness and efficiency of the proposed robust DRL-based EH-RIS system, which not only outperform the existing state-of-the-art solutions but also approach to the performance of the exhaustive search method.

DeepHistoNet: A robust deep-learning model for the classification of hepatocellular, lung, and colon carcinoma.

In recent days, non-communicable diseases (NCDs) require more attention since they require specialized infrastructure for treatment. As per the cancer population registry estimate, nearly 800,000 new cancer cases will be detected yearly. The statistics alarm the need for early cancer detection and diagnosis. Cancer identification can be made either through manual efforts or by computer-aided algorithms. Manual efforts-based cancer detection is labor intensive and also offers more time complexity. In contrast, computer-aided algorithms offer feasibility in reducing time and manual efforts. With the motivation to develop a computer-aided diagnosis system for NCD, we developed a cancer detection methodology. In the present article, a deep learning (DL)-based cancer identification model is developed. In DL-based architectures, the features are generally extracted using convolutional neural networks. The proposed attention-guided, densely connected residual, and dilated convolution deep neural network called DeepHistoNet acquire precise patterns for classification. Experimentation has been carried out on Kasturba Medical College (KMC), TCGA-LIHC, and LC25000 datasets to prove the robustness of the model. Performance evaluation metrics like F1-score, sensitivity, specificity, recall, and accuracy validate the experimentation. Experimental results demonstrate that the proposed DeepHistoNet model outperforms the other state-of-the-art methods. The proposed model has been able to classify the KMC liver dataset with 97.1% accuracy and 0.9867 value of area under the curve-receiver operating characteristic curve (AUC-ROC), which is the best result obtained compared to the state-of-the-art techniques. The performance of the DeepHistoNet has been even better on the LC25000 dataset. On the LC25000 dataset, the proposed model achieved 99.8% classification accuracy. To our knowledge, DeepHistoNet is a novel approach for multiple histopathological image classification. RESEARCH HIGHLIGHTS: A novel robust DL model is proposed for histopathological image carcinoma classification. The precise patterns for accurate classification are extracted using dense cross-connected residual blocks. Spatial attention is provided to the network so that the spatial information is not lost during the feature extraction. DeepHistoNet is trained and evaluated on the liver, lung, and colon histopathology datasets to demonstrate its resilience. The results are promising and outperform state-of-the-art techniques. The proposed methodology has obtained the AUC-ROC value of 0.9867 with a classification accuracy of 97.1% on the KMC dataset. The proposed DeepHistoNet has classified the LC25000 dataset with 99.8% accuracy. The results are the best obtained till date.

Correction: Beyond accuracy and precision: a robust deep learning framework to enhance the resilience of face mask detection models against adversarial attacks

Correction: Beyond accuracy and precision: a robust deep learning framework to enhance the resilience of face mask detection models against adversarial attacks

A robust deep learning workflow to predict CD8 + T-cell epitopes

BackgroundT-cells play a crucial role in the adaptive immune system by triggering responses against cancer cells and pathogens, while maintaining tolerance against self-antigens, which has sparked interest in the development of various T-cell-focused immunotherapies. However, the identification of antigens recognised by T-cells is low-throughput and laborious. To overcome some of these limitations, computational methods for predicting CD8 + T-cell epitopes have emerged. Despite recent developments, most immunogenicity algorithms struggle to learn features of peptide immunogenicity from small datasets, suffer from HLA bias and are unable to reliably predict pathology-specific CD8 + T-cell epitopes.MethodsWe developed TRAP (T-cell recognition potential of HLA-I presented peptides), a robust deep learning workflow for predicting CD8 + T-cell epitopes from MHC-I presented pathogenic and self-peptides. TRAP uses transfer learning, deep learning architecture and MHC binding information to make context-specific predictions of CD8 + T-cell epitopes. TRAP also detects low-confidence predictions for peptides that differ significantly from those in the training datasets to abstain from making incorrect predictions. To estimate the immunogenicity of pathogenic peptides with low-confidence predictions, we further developed a novel metric, RSAT (relative similarity to autoantigens and tumour-associated antigens), as a complementary to ‘dissimilarity to self’ from cancer studies.ResultsTRAP was used to identify epitopes from glioblastoma patients as well as SARS-CoV-2 peptides, and it outperformed other algorithms in both cancer and pathogenic settings. TRAP was especially effective at extracting immunogenicity-associated properties from restricted data of emerging pathogens and translating them onto related species, as well as minimising the loss of likely epitopes in imbalanced datasets. We also demonstrated that the novel metric termed RSAT was able to estimate immunogenic of pathogenic peptides of various lengths and species. TRAP implementation is available at: https://github.com/ChloeHJ/TRAP.ConclusionsThis study presents a novel computational workflow for accurately predicting CD8 + T-cell epitopes to foster a better understanding of antigen-specific T-cell response and the development of effective clinical therapeutics.

A highly robust deep learning technique for overlap detection using audio fingerprinting

A highly robust deep learning technique for overlap detection using audio fingerprinting

Robust Deep Learning Based Framework for Detecting Cyber Attacks from Abnormal Network Traffic

The internet's recent rapid growth and expansion have raised concerns about cyberattacks, which are constantly evolving and changing. As a result, a robust intrusion detection system was needed to safeguard data. One of the most effective ways to meet this problem was by creating the artificial intelligence subfields of machine learning and deep learning models. Network integration is frequently used to enable remote management, monitoring, and reporting for cyber-physical systems (CPS). This work addresses the primary assault categories such as Denial of Services(DoS), Probe, User to Root(U2R) and Root to Local(R2L) attacks. As a result, we provide a novel Recurrent Neural Networks (RNN) cyberattack detection framework that combines AI and ML techniques. To evaluate the developed system, we employed the Network Security Laboratory-Knowledge Discovery Databases (NSL-KDD), which covered all critical threats. We used normalisation to eliminate mistakes and duplicated data before pre-processing the data. Linear Discriminant Analysis(LDA) is used to extract the characteristics. The fundamental rationale for choosing RNN-LDA for this study is that it is particularly efficient at tackling sequence issues, time series prediction, text generation, machine translation, picture descriptions, handwriting recognition, and other tasks. The proposed model RNN-LDA is used to learn time-ordered sequences of network flow traffic and assess its performance in detecting abnormal behaviour. According to the results of the experiments, the framework is more effective than traditional tactics at ensuring high levels of privacy. Additionally, the framework beats current detection techniques in terms of detection rate, false positive rate, and processing time.

THPoseLite, a Lightweight Neural Network for Detecting Pose in Thermal Images

Nowadays, Smart Environments (SEs) enable the monitoring of people with physical disabilities by incorporating activity recognition. Thermal cameras are being incorporated as they preserve privacy. Some Deep Learning (DL) solutions use the pose of the users because it removes external noise. Although there are robust DL solutions in the visible spectrum, they fail in the thermal domain. Thus, we propose THPoseLite (Thermal Human Pose Lite), a Convolutional Neural Network (CNN) based on MobileNetV2 that extracts pose from Thermal Images (TIs). In a novel way, an auto-labeling approach has been developed. It includes a background removal using an optical flow estimator. It also integrates Blazepose (a pose estimator for Visible spectrum images (VSIs)) to obtain the poses in the pre-processed TIs. Results show that the pre-processing increases the percentage of detected poses by Blazepose from 19.55% to 76.85%. This allows the recording of Human Pose Estimation (HPE) datasets in the visible spectrum without requiring visible spectrum cameras or manually annotating datasets. Furthermore, THPoseLite has been embedded in an Internet of Things (IoT) device incorporating an edge Tensor Processing Unit (TPU) accelerator, which can process TIs recorded at 9 Frames Per Second (FPS) in real-time (12.28 FPS). It requires fewer than 6W of energy to run. It has been achieved using model quantization, decreasing the accuracy in estimating the poses by only 1%. The MSE of MobileNetV2 in test images is 35.48, obtaining accurate poses in 21% of the images that Blazepose is not able to detect any pose.

Noise robust Zernike phase retrieval via learning based algorithm only with 2-step phase shift measurements.

We present a noise robust deep learning based aberration analysis method using 2-step phase shift measurement data. We first propose a realistic aberration pattern generation method to synthesize a sufficient amount of real-world-like aberration patterns for training a deep neural network by exploiting the asymptotic statistical distribution parameters of the real-world Zernike coefficients extracted from a finite number of experimentally measured real-world aberration patterns. As a result, we generate a real-world-like synthetic dataset of 200,000 different aberrations from 15 sets of real-world aberration patterns obtained by a Michelson interferometer under a variety of measurement conditions using the 4-step derivative fitting method together with the exploitation of the Gaussian density estimation. We then train the deep neural network with the real-world-like synthetic dataset, using two types of network architectures, GoogLeNet and ResNet101. By applying the proposed learning based 2-step aberration analysis method to the analysis of numerically generated aberrations formed under 100 different conditions, we verify that the proposed 2-step method can clearly outperform the existing 4-step iterative methods based on 4-step measurements, including the derivative fitting, transport of intensity equation (TIE), and robust TIE methods, in terms of noise robustness, root mean square error (RMSE), and inference time. By applying the proposed 2-step method to the analysis of the real-world aberrations experimentally obtained under a variety of measurement conditions, we also verify that the proposed 2-step method achieves compatible performance in terms of the RMSE between the reconstructed and measured aberration patterns, and also exhibits qualitative superiority in terms of reconstructing more realistic fringe patterns and phase distributions compared to the existing 4-step iterative methods. Since the proposed 2-step method can be extended to an even more general analysis of aberrations of any higher order, we expect that it will be able to provide a practical way for comprehensive aberration analysis and that further studies will extend its usefulness and improve its operational performance in terms of algorithm compactness, noise robustness, and computational speed.

Content-Based Image Retrieval using Hard Voting Ensemble Method of Inception, Xception, and Mobilenet Architectures

Advancements in internet accessibility and the affordability of digital picture sensors have led to the proliferation of extensive image databases utilized across a multitude of applications. Addressing the semantic gap between low-level attributes and human visual perception has become pivotal in refining Content-Based Image Retrieval (CBIR) methodologies, especially within this context. As this field is intensely researched, numerous efficient algorithms for CBIR systems have surfaced, precipitating significant progress in the artificial intelligence field. In this study, we propose employing a hard voting ensemble approach on features derived from three robust deep learning architectures: Inception, Exception, and Mobilenet. This is aimed at bridging the divide between low-level image features and human visual perception. The Euclidean method is adopted to determine the similarity metric between the query image and the features database. The outcome was a noticeable improvement in image retrieval accuracy. We applied our approach to a practical dataset named CBIR 50, which encompasses categories such as mobile phones, cars, cameras, and cats. The effectiveness of our method was thereby validated. Our approach outshone existing CBIR algorithms with superior accuracy (ACC), precision (PREC), recall (REC), and F1-score (F1-S), proving to be a noteworthy addition to the field of CBIR. Our proposed methodology could be potentially extended to various other sectors, including medical imaging and surveillance systems, where image retrieval accuracy is of paramount importance.

Real-time gait biometrics for surveillance applications: A review

Deep learning (DL) pipelines have evolved for over a decade now and are efficient at solving many challenging problems of image and signal processing applications. Designing deep learning pipelines for a particular application requires a good understanding of deep learning and various intermediate layers available. To develop a DL pipeline, one uses available dataset(s) suitable for an application, and the pipeline is refined by iterating over intermediate layers. A large amount of time and extensive thinking goes into these selections and validating the performance of each configuration. Thus, it is hard to choose the correct and robust DL pipeline that performs well on all relevant datasets. This review aims to aid researchers in understanding different gait sensing technologies and provide foundational knowledge of the deep learning concepts for faster solutions for a given problem. Gait recognition is more recent since it hasn’t yet been used in a real-world situation. This article provides a comprehensive overview of gait biometrics suited to real-time surveillance applications. All the important parameters of deep learning pipelines are explained, along with their selection and implication for a given problem. Authors have reviewed important research articles recently on deep learning models and how these perform across different application datasets. The benefits and drawbacks of the approaches are elucidated to help arrive at the optimized pipeline derived from a fusion of available pipelines to achieve faster but accurate results for a given problem.