Srpr net improves instance segmentation using semantic and relational prompts
SRPR-Net: Semantic and Relational Prompt Refinement for Automated SAM-based Instance Segmentation
Computer Vision and Pattern Recognition
Summary
Instance segmentation involves identifying and separating objects in images. The authors found that current automated methods often miss important details about each object and how objects relate to one another. They created SRPR-Net, which improves accuracy by refining prompts that guide object detection using both language understanding and the relationships between objects. This enhanced approach helps better locate and segment objects in images, as shown by tests on standard datasets.
What this means in practice
- •For computer vision engineers: Integrate context-aware prompt refinement to improve object instance detection and segmentation accuracy in image analysis systems.
- •For robotics developers: Enhance robot perception modules with relational prompt methods for better understanding of object boundaries and interactions in complex scenes.
Authors
Lufei Liu, Guojie Li, Suncheng Xiang, Fan Zhang
Abstract
Instance segmentation is a fundamental computer vision task with diverse real-world applications. Recently, prompt-driven foundation models have shown promising generalization. However, automated prompting remains limited by insufficient semantic guidance and inter-instance modeling. To address this challenge, we propose a novel architecture, named Semantic Relational Prompt Refinement Network (SRPR-Net), for automated SAM-based instance segmentation. A sequential prompt refinement mechanism is introduced to enrich detector geometry with visual-language semantics and then incorporate same-image instance dependencies, enabling context-aware box adjustment before SAM segmentation. Experiments on multiple standard benchmarks demonstrate that SRPR-Net achieves consistent improvements in segmentation performance over existing state-of-the-art approaches. The code is publicly available at https://github.com/JeremyXSC/SRPR-Net.