As intelligent educational technologies advance rapidly, recognizing classroom behavior has become essential for enhancing teaching effectiveness and enabling personalized learning. Traditional recognition methods, however, encounter issues like sparse datasets, occlusions, and challenges in identifying small objects within complex classroom settings. This study introduces an enhanced YOLOv8 model, referred to as DLKAS-YOLOv8, which integrates the C2f-Faster module. First, the C2f-Faster module is incorporated to replace YOLOv8’s original Bottleneck module, aiming to minimize redundant computations and boost feature extraction capabilities. Additionally, the Deformable Large Kernel Attention (DLKA) mechanism is applied to capture both detailed and broader contextual information, enhancing the model's versatility in challenging scenes. The Slide loss function is also employed to improve the model's handling of imbalanced data. Through comparative testing on a classroom behavior dataset, evaluated using mean Average Precision (mAP), Precision, and Recall metrics, DLKAS-YOLOv8 demonstrates superior performance in detection accuracy and processing speed, particularly for small and overlapping objects. Future research will integrate object tracking and facial recognition to further optimize the model's real-time capabilities and accuracy.

Jiangfeng Sun, Songyan Li, Jianchun Zhang. Classroom Behavior Recognition and Research Based on DLKAS-YOLO8n. Academic Journal of Computing & Information Science (2024), Vol. 7, Issue 11: 109-116. https://doi.org/10.25236/AJCIS.2024.071115.

[1] Chen Li. Research on "Students' Classroom Problem Behaviors" from the Perspective of Senior High School Teachers [D]. Nanchang University.,2024.DOI:10.27232/d.cnki.gnchu.2024.001111.

[2] Zhao Ke. Methods and Implementations for Human Skeleton Behavior Recognition in Classroom Scenarios [D]. Shenyang University,2024.DOI:10.27692/d.cnki.gsydx.2024.000125.

[3] Zhou Ye. Research on the Detection of Primary School Students' Classroom Behaviors Based on Faster R-CNN [D]. Sichuan Normal University,2021.DOI:10.27347/d.cnki.gssdu.2021.000962.

[4] GUO M H, LU C Z, LIU Z N, et al. Visual attention network[J]. Computational Visual Media, 2023, 9(4): 733-752. 

[5] HOIEM D, DIVVALA S K, Hays J H. Pascal VOC 2008 challenge[J]. World Literature Today, 2009, 24(1): 1-4. 

[6] PADILLA R, NETTO S L, Da Silva E A B. A survey on performance metrics for object-detection algorithms [C]/ 2020 International Conference on Systems, Signals and Image Processing (IWSSIP), Niteroi, Brazil, 2020: IEEE, 2020: 237-242. 

[7] ZHAO Y, KV W, XU S, et al. Detrs beat yolos on real-time object detection[C]//Proceedings of the IEEE/ CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, WA, 2024: IEEE, 2024: 16965- 16974. 

[8] TAN M, PANG R, Le Q V. Efficientdet: scalable and efficient object detection[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, WA, 2020: IEEE, 2020: 10781-10790. 

[9] MUSHTAQ F, RAMESH K, DESHMUKH S, et al. Nuts&bolts: YOLO-v5 and image processing based component identification system[J]. Engineering Applications of Artificial Intelligence, 2023, 118: 105665

[10] ATREY J, REGUNATHAN R, RAJASEKARAN R. Realworld application of face mask detection system using YOLOv6 [J]. International Journal of Critical Infrastructures, 2024, 20(3): 216-240. 

[11] PADILLA R, NETTO S L, Da Silva E A B. A survey on performance metrics for object-detection algorithms [C]/ 2020 International Conference on Systems, Signals and Image Processing (IWSSIP), Niteroi, Brazil, 2020: IEEE, 2020: 237-242. 

[12] ZHAO Y, KV W, XU S, et al. Detrs beat yolos on real-time object detection[C]//Proceedings of the IEEE/ CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, WA, 2024: IEEE, 2024: 16965- 16974. 

[13] TAN M, PANG R, Le Q V. Efficientdet: scalable and efficient object detection[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, WA, 2020: IEEE, 2020: 10781-10790. 

[14] Cheng Wenhao. Research on Classroom Student Detection and Behavior Recognition Based on YOLOv5 Deep Learning [D]. Chang'an University,2023.DOI:10.26976/d.cnki.gchau.2023.002213. 

[15] Gao Fei. Research on Human Pose Estimation Based on Multi-Network Fusion [D]. Dalian Jiaotong University,2024.DOI:10.26990/d.cnki.gsltc.2024.000012.

[16] JOCHE G, CHAURASIA A. (2023). Ultralytics/YOLOv8 in PyTorch. URL: https://github. com/ ultralytics/ultralytic

[17] Chen Yuling. Research on Small Target Feature Enhancement and Detection Methods Based on Convolutional Neural Network [D]. Southwest University of Science and Technology, 2024. DOI:10.27415/d.cnki.gxngc.2024.000164.

[18] Zeng Yuqi, Liu Bo, Zhong Baichang, et al. Student Classroom Behavior Detection Algorithm Based on Improved YOLOv8 in the Context of Smart Education [J]. Computer Engineering,2024,50(09):344-355. DOI:10.19678/j.issn.1000-3428.0069597.

[19] Azad R, Niggemeier L, Hüttemann M, et al. Beyond self-attention: Deformable large kernel attention for medical image segmentation [C]. Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision. 2024: 1287-1297.

[20] YU Z,HUANG H,CHEN W,et al.YOLO-faceV2:a scale and occlusion aware face detector[J].Pattern Recognition,2024,155:110714.

[21] WANG W, DAI J,CHEN Z, et al. Intern image: Exploring large-scale vision foundation models with deformable convolutions[C]// 2023 IEEE Conference on Computer Vision and Pattern Recognition. Vancouver, Canada: IEEE, 2023: 14408-14419.

[22] Cui Shaoguo, Xu Song, Wang Mingyang, et al. Research Progress of Deep Learning Knowledge Tracing for Intelligent Education [J/OL]. Computer Engineering,1-18[2024-10-27]. https://doi.org/10.19678/j.issn.1000-3428.0069653.

[23] Liu Genghuan, Zeng Xiangjin, Dou Jiazhen, et al. Research Progress of Small Target Detection Technology Based on Deep Learning (Invited) [J]. Infrared and Laser Engineering,2024,53(09):196-228.

[24] Chen Chen, Bao Wenxing, Chen Xu, et al. Student Classroom Behavior Recognition Algorithm Based on Improved YOLOv8: DMS-YOLOv8 [J/OL]. Computer Engineering and Applications,1-17[2024-10-27].http://kns.cnki.net/kcms/detail/11.2127.tp.20240913.1601.008.html.

[25] Chen J, Kao S, He H, et al. Run, Don't Walk: Chasing Higher FLOPS for Faster Neural Networks[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2023:12021-12031.