The academic level of students is the core issue that schools and parents pay attention to. There is a close relationship between students' academic level and students' attention status in class. Students who can maintain their concentration for a long time tend to have higher academic level. Therefore, it is very necessary to improve students' attention level in class. To this end, the attention of the students needs to be detected first. Computer vision algorithms typically employ third-person image or video data. Due to the large number of students in the classroom, the algorithm is difficult to extract small objects and occluded objects, so it cannot accurately detect students' attention. The attention detection model proposed in this paper included three modules: gaze point estimation, gaze target recognition and attention level analysis. The gaze estimation module used a deep learning algorithm that combines saliency detection with attention shifting. In order to improve the accuracy of the algorithm, this paper tried three attention mechanisms, namely spatial attention, channel attention and mixed attention. The specific approach was to embed the attention module after the convolution block of the Convolutional Neural Network (CNN). In order to achieve the goal of obtaining students' classroom attention data in batches and accurately, this paper proposed a first-person video-based student classroom attention detection method. Compared with the third-person video, the first-person video had the advantages of one-to-one correspondence between the video and the students, and the video content was consistent with the gaze behavior, so it could make the students' attention detection more accurate. The experiments in this paper showed that the sum of the ratio of student 1's fixation to learning-related goals was 0.51, while that of student 2 was 0.56. Therefore, student 2's attention state was more concentrated, and the final algorithm model was better than the baseline.

Lian Xue. Classroom Attention Detection Based on Computer Vision and Artificial Intelligence. International Journal of Big Data Intelligent Technology (2025), Vol. 6, Issue 1: 33-46. https://doi.org/10.38007/IJBDIT.2025.060103.

[1] Li S Y, Xu B G, Fu H, Tao X M, Chi Z R. A two-scale attention model for intelligent evaluation of yarn surface qualities with computer vision [J]. Journal of the Textile Institute, 2018, 109 (6): 798-812. 

[2] Oord S, Boyer B E, Dyck L V, Mackay K J, Meyer H D, Baeyens D. A Randomized Controlled Study of a Cognitive Behavioral Planning Intervention for College Students With ADHD: An Effectiveness Study in Student Counseling Services in Flanders: [J]. Journal of Attention Disorders, 2020, 24 (6): 849-862. 

[3] Norazman N, Che-Ani A I, Ja'Afar N H, Khoiry M A. Standard compliance and suitability of classroom capacity in secondary school buildings [J]. Journal of Facilities Management, 2019, 17 (3): 238-248. 

[4] maharmeh, lina, mahmoud. Reducing The Rate of Behavioral Problems for Students with ASD & ADHD using the Techniques of FBA [J]. International Journal for Research in Education, 2019, 43 (2): 9-9. 

[5] Abate A F, Cascone L, Nappi M, Narducci F, Passero I. Attention monitoring for synchronous distance learning [J]. Future Generation Computer Systems, 2021, 125 (4): 774-784. 

[6]Caroline, Guardino, Katrina, W., Hall, Erin, et al. Teacher and student perceptions of an outdoor classroom [J]. Journal of Outdoor and Environmental Education, 2019, 22 (2): 113–126. 

[7] Fu K, Jin J, Cui R, Fei S, Zhang C. Aligning Where to See and What to Tell: Image Captioning with Region-Based Attention and Scene-Specific Contexts [J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2017, 39 (12): 2321-2334. 

[8] Attigodu G, Berthommier F, Nahorna O, Schwartz J L, Attigoduchandrashekara G, Olhanahorna J L, et al. effect of context, rebinding and noise, on audiovisual speech fusion. 14th annual conference of the international speech communication association [J]. Exp Brain Res, 2017, 184 (1): 39-52. 

[9] Liu S, Huang D, Wang Y. Pay Attention to Them: Deep Reinforcement Learning-Based Cascade Object Detection [J]. IEEE Transactions on Neural Networks and Learning Systems, 2020, 31 (7): 2544-2556. 

[10] Sangeroki B A, Cenggoro T W. A Fast and Accurate Model of Thoracic Disease Detection by Integrating Attention Mechanism to a Lightweight Convolutional Neural Network [J]. Procedia Computer Science, 2021, 179 (11): 112-118. 

[11] Luo X, Hu H. Selected and refined local attention module for object detection [J]. Electronics Letters, 2020, 56 (14): 712-714. 

[12] Wang Y, Gu X. Using of Attention for Scene Text Detection [J]. Journal of Computer-Aided Design & Computer Graphics, 2022, 33 (12): 1908-1915. 

[13] Zhou M, Zou Z, Shi Z, Zeng W J, Gui J. Local Attention Networks for Occluded Airplane Detection in Remote Sensing Images [J]. IEEE Geoscience and Remote Sensing Letters, 2020, 17 (3): 381-385. 

[14] Liang Y, Qin G, Sun M, Yan J, Jiang H. MAFNet: Multi-style attention fusion network for salient object detection [J]. Neurocomputing, 2021, 422 (2): 22-33. 

[15] Avoliot B J, Gardner W L. Authentic leadership development: Getting to the root of positive forms of leadership [J]. IEEE Engineering Management Review, 2017, 16 (3): 315-338. 

[16] Baglio K J. Student Motivation in the Latin Classroom [J]. Journal of Classics Teaching, 2022, 23 (45): 75-78. 

[17] Chen J, Wan L, Zhu J, Xu G, Deng M. Multi-Scale Spatial and Channel-wise Attention for Improving Object Detection in Remote Sensing Imagery [J]. IEEE Geoscience and Remote Sensing Letters, 2020, 17 (4): 681-685. 

[18] Kempf E, Manconi A, Spalt O. Distracted Shareholders and Corporate Actions [J]. Review of Financial Studies, 2017, 30 (5): 1660-1695. 

[19] Ren S, He K, Girshick R, Sun J. Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks [J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2017, 39 (6): 1137-1149. 

[20] Zhao Z, Bao Z, Zhang Z, Cummins N, Sun S, Wang H, et al. Self-attention transfer networks for speech emotion recognition [J]. Virtual Reality & Intelligent Hardware, 2021, 3 (1): 43-54.