Human Action Recognition Using Computer Vision: A Deep Learning Approach / (Record no. 610816)
[ view plain ]
| 000 -LEADER | |
|---|---|
| fixed length control field | 02604nam a22001577a 4500 |
| 082 ## - DEWEY DECIMAL CLASSIFICATION NUMBER | |
| Classification number | 610 |
| 100 ## - MAIN ENTRY--PERSONAL NAME | |
| Personal name | Hussain, Fatima |
| 245 ## - TITLE STATEMENT | |
| Title | Human Action Recognition Using Computer Vision: A Deep Learning Approach / |
| Statement of responsibility, etc. | Fatima Hussain |
| 264 ## - PRODUCTION, PUBLICATION, DISTRIBUTION, MANUFACTURE, AND COPYRIGHT NOTICE | |
| Place of production, publication, distribution, manufacture | Islamabad : |
| Name of producer, publisher, distributor, manufacturer | SMME- NUST; |
| Date of production, publication, distribution, manufacture, or copyright notice | 2024. |
| 300 ## - PHYSICAL DESCRIPTION | |
| Extent | 90p. |
| Other physical details | Soft Copy |
| Dimensions | 30cm |
| 500 ## - GENERAL NOTE | |
| General note | Human action recognition (HAR) is always an enthralling topic because it facilitates the<br/>identification of activity from the video's sequence. Applications for Human action<br/>recognition is numerous including surveillance, sport analysis, suspicious activity<br/>recognition and healthcare. Human activity recognition is hampered by poor resolution<br/>cameras, extreme weather, and similar colors for both the subject and the object, as well<br/>as by intraclass human activity such as walking and jogging. Currently available<br/>approaches i.e. transformer based models, expanded datasets and improved temporal<br/>modelling techniques such as attention mechanisms and LSTMs remove the background<br/>noise from the final layers but the accuracy of correctly identifying actions is reduced and<br/>address intraclass resemblance in human action classification to some extent. These<br/>advancements improve the capabilities of action recognition systems but completely<br/>resolving intraclass resemblance is a challenging task. Therefore, there is a growing need<br/>for improved computer vision-based surveillance systems. A hybrid approach called<br/>"Human Action Recognition using Deep Learning and Hybrid Evolutionary Techniques"<br/>is proposed to address these issues. It consists of following main steps: preprocessing i.e.<br/>contrast enhancement, data augmentation, customized models based on residual block<br/>architecture, training Residual Block2 and Residual Block3 models, feature extraction<br/>and testing, features fusion, feature selection using Binary Chimp optimization and<br/>classification. To enhance interpretability, transparency and trust in machine learning<br/>models, Grad-CAM and LIME are applied. Both these techniques provide visual display<br/>of important regions in imaging. Grad-CAM gave heatmaps and LIME produced<br/>highlighted regions on original images. Our suggested methodology achieves state-ofthe-art accuracy on the UT Interaction dataset of Action Recognition with 94% Accuracy.<br/>This emphasizes how well the proposed technique works to improve the classification of<br/>human actions. |
| 650 ## - SUBJECT ADDED ENTRY--TOPICAL TERM | |
| Topical term or geographic name entry element | MS Biomedical Sciences (BMS) |
| 700 ## - ADDED ENTRY--PERSONAL NAME | |
| Personal name | Supervisor : Prof. Dr. Javaid Iqbal |
| 856 ## - ELECTRONIC LOCATION AND ACCESS | |
| Uniform Resource Identifier | <a href="http://10.250.8.41:8080/xmlui/handle/123456789/45274">http://10.250.8.41:8080/xmlui/handle/123456789/45274</a> |
| 942 ## - ADDED ENTRY ELEMENTS (KOHA) | |
| Source of classification or shelving scheme | |
| Koha item type | Thesis |
| Withdrawn status | Permanent Location | Current Location | Shelving location | Date acquired | Full call number | Barcode | Koha item type |
|---|---|---|---|---|---|---|---|
| School of Mechanical & Manufacturing Engineering (SMME) | School of Mechanical & Manufacturing Engineering (SMME) | E-Books | 08/08/2024 | 610 | SMME-TH-1040 | Thesis |