Reasoning segmentation (RS) interprets implicit textual instructions to accurately segment target regions. This reasoning capability transforms ambiguous non-expert queries into precise pixel-level masks, thereby enabling downstream tasks like area measurement and density analysis with a level of precision unattainable by detection methods. However, existing RS models are not tailored for agriculture and lack domain-specific knowledge, which poses challenges in handling similar pest appearances and small target scales. To bridge this gap, we introduce a fine-grained pest RS task with two subtasks: Pest Discriminative Referring Expression Segmentation (PDRES) and Pest Exclusion Reasoning Segmentation (PERS). Based on this, we propose PestScope, which integrates vision, language, and reasoning for fine-grained pest segmentation. To tackle the exclusion of small non-target pests, we introduce a dedicated [NON] token alongside the standard [SEG] token for target pests. This guides the model to prioritize small target pests and suppress non-target background regions. To further address pest similarity, we propose an Exclusivity Suppression Loss, applying differentiated supervision to [SEG] and [NON] tokens to better separate target and non-target pests. Additionally, we develop an automated dataset construction pipeline to address the scarcity of fine-grained, difficulty-controllable pest RS datasets. It produces 45k and 27.6k image-text-mask samples for the PDRES and PERS tasks, respectively, covering 18 pest categories. Experiments show that in small and similar pest scenarios, integrating PestScope into mainstream models improves average gIoU by 4.28% on PDRES and 6.49% on PERS. For unseen pest categories, gIoU increases by 21.72% and 8.66%, respectively, demonstrating strong generalization. Code and datasets will be available at: https://github.com/aluodaydayup/PestScope.

Abstract In recent years, artificial intelligence has been applied in many fields, and education has attracted more and more attention. More and more behaviour detection and recognition algorithms are applied in the field of education. Students' attention in class is the key to improving the quality of teaching, and classroom behavior is a direct manifestation of students' attention. In view of the problem that the accuracy of students' classroom behavior recognition is generally low, we apply deep learning to multi‐view behavior detection, which can detect and recognize behaviors from different perspectives, to evaluate students' classroom attention. First, an improved detection model based on YOLOv5 is proposed, which improves the CBL module throughout the entire network to optimize the model and uses SIoU as the loss function to improve the convergence speed of the prediction box. Second, a quantitative evaluation standard for students' classroom attention is established and then training and verification are conducted by collecting multi‐view classroom datasets. Finally, the environment variation in the training model phase is increased to make the model have better generalization ability. Experiments demonstrate that our method can effectively identify and detect students' behaviours in the classroom from different angles, and it has good robustness and feature extraction capabilities.