Enhancing real-time instance segmentation for plant disease detection with improved YOLOv8-Seg algorithm

Abstract

With widespread uses in areas as diverse as traffic analysis and medical imaging, picture segmentation is a basic problem in computer vision. Instance segmentation, which combines object recognition with segmentation, is a powerful tool for item identification and exact delineation. Using the Tomato Leaf disease dataset as an example, this research delves into the topic of segmentation training by capitalizing on the simplicity of enhanced YOLOv8-Seg models. Tomato leaf disease are the focus of this instance-segmentation dataset, which seeks to resolve the pressing problem of agricultural difficulties. One instance segmentation networks, YOLOv8n-Seg is presented and compared in this article for the purpose of Tomato leaf disease identification. The models are tested in difficult situations to see how well they can detect and separate garbage occurrences. Results show that enhanced YOLOv8-Seg is useful for agriculture by accurately segmenting instances of tomato leaf disease detection.