Human action recognition: a construction of codebook by discriminative features selection approach

Siddiqui, Samra; Khan, Muhammad Attique; Bashir, Khalid; Sharif, Muhammad; Azam, Faisal; Javed, Muhammad Awais

doi:10.1504/ijapr.2018.094815

Cited by 20 publications

(3 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Human activity recognition is an active research field 6 , in which traditional methods mainly rely on a single modality and can be broadly categorized into three types: (1) Handcrafted representation-based methods [7][8][9][10][11][12][13][14] .…”

Section: Unimodal Activity Recognitionmentioning

confidence: 99%

Multiscale knowledge distillation with attention based fusion for robust human activity recognition

Yuan,

Yang,

Ning

et al. 2024

Sci Rep

View full text Add to dashboard Cite

Knowledge distillation is an effective approach for training robust multi-modal machine learning models when synchronous multimodal data are unavailable. However, traditional knowledge distillation techniques have limitations in comprehensively transferring knowledge across modalities and models. This paper proposes a multiscale knowledge distillation framework to address these limitations. Specifically, we introduce a multiscale semantic graph mapping (SGM) loss function to enable more comprehensive knowledge transfer between teacher and student networks at multiple feature scales. We also design a fusion and tuning (FT) module to fully utilize correlations within and between different data types of the same modality when training teacher networks. Furthermore, we adopt transformer-based backbones to improve feature learning compared to traditional convolutional neural networks. We apply the proposed techniques to multimodal human activity recognition and compared with the baseline method, it improved by 2.31% and 0.29% on the MMAct and UTD-MHAD datasets. Ablation studies validate the necessity of each component.

show abstract

Section: Unimodal Activity Recognitionmentioning

confidence: 99%

Multiscale knowledge distillation with attention based fusion for robust human activity recognition

Yuan,

Yang,

Ning

et al. 2024

Sci Rep

View full text Add to dashboard Cite

show abstract

“…In this approach, the k-nearest neighbor method was used as a fitness method to determine the performance of the optimized feature set. Siddiqui et al (2018) presented a codebook-based feature selection approach that includes models of visual vocabulary learning, quantization of features based on learned visual vocabulary, and representation of images by using the frequency of visual words. In the final module, activity classification was conducted using the support vector machine algorithm.…”

Section: Related Workmentioning

confidence: 99%

Quantum behaved Intelligent Variant of Gravitational Search Algorithm with Deep Neural Networks for Human Activity Recognition

Jindal

Sachdeva

Kushwaha

2022

KJS publishes peer-review articles in Mathematics, Computer Sci

View full text Add to dashboard Cite

Human activity recognition (HAR) encompasses the detection of daily routine activities to advance usability in detecting crime and preventing dangerous activities. The recognition of activities from videos and image sequences with higher exactitude is a major challenge due to system complexities. The efficient feature optimization approach can reduce system complexities by removing ineffective features, which also improves the activity recognition performance. This research work presents a novel quantum behaved intelligent gravitational search algorithm to optimize the features for human activity recognition. The proposed intelligent variant is termed as INQGSA, which optimizes the features by using the advantageous attributes of quantum computing (QC) and intelligent gravitational search algorithm (INGSA). In INQGSA, the intelligent factor avoids the trapping of mass agents in later iterations by using the information of the best and worst agents to update the position of agents. The addition of quantum computing based attributes (such as quantum bits, their superposition, and quantum gates, etc.) ensures a better diversity of discrete optimized features. To analyze the superiority of INQGSA, the feature optimization is also conducted with the gravitational search algorithm (GSA) and the quantum-inspired binary gravitational search algorithm (QBGSA). Finally, the optimized selected features are utilized by the deep neural networks (DNN) of ResNet-50V2 and ResNet-101V2 for the classification of activities. The activity recognition experiments are conducted on the UCF101 and HMDB51 datasets. The performance comparison of the proposed HAR system with state-of-the-art techniques signifies that the proposed system is superior and effective in detecting the different activities.

show abstract

“…Further, the main limitation of this process is that if a human silhouette is incorrectly detected, irrelevant features are extracted [23,24]. Researchers have tried to resolve these issues using feature reduction techniques such as principle component analysis (PCA) [25,26], linear discriminant analysis (LDA) [27], and improved PCA [28], among others [29,30].…”

Section: Introductionmentioning

confidence: 99%

Human Gait Recognition: A Single Stream Optimal Deep Learning Features Fusion

Saleem

Khan

Alhaisoni

et al. 2021

Sensors

Self Cite

View full text Add to dashboard Cite

Human Gait Recognition (HGR) is a biometric technique that has been utilized for security purposes for the last decade. The performance of gait recognition can be influenced by various factors such as wearing clothes, carrying a bag, and the walking surfaces. Furthermore, identification from differing views is a significant difficulty in HGR. Many techniques have been introduced in the literature for HGR using conventional and deep learning techniques. However, the traditional methods are not suitable for large datasets. Therefore, a new framework is proposed for human gait recognition using deep learning and best feature selection. The proposed framework includes data augmentation, feature extraction, feature selection, feature fusion, and classification. In the augmentation step, three flip operations were used. In the feature extraction step, two pre-trained models were employed, Inception-ResNet-V2 and NASNet Mobile. Both models were fine-tuned and trained using transfer learning on the CASIA B gait dataset. The features of the selected deep models were optimized using a modified three-step whale optimization algorithm and the best features were chosen. The selected best features were fused using the modified mean absolute deviation extended serial fusion (MDeSF) approach. Then, the final classification was performed using several classification algorithms. The experimental process was conducted on the entire CASIA B dataset and achieved an average accuracy of 89.0. Comparison with existing techniques showed an improvement in accuracy, recall rate, and computational time.

show abstract

Human action recognition: a construction of codebook by discriminative features selection approach

Cited by 20 publications

References 0 publications

Multiscale knowledge distillation with attention based fusion for robust human activity recognition

Multiscale knowledge distillation with attention based fusion for robust human activity recognition

Quantum behaved Intelligent Variant of Gravitational Search Algorithm with Deep Neural Networks for Human Activity Recognition

Human Gait Recognition: A Single Stream Optimal Deep Learning Features Fusion

Contact Info

Product

Resources

About