LLM-Seg: Bridging Image Segmentation and Large Language Model Reasoning
This work addresses the challenge of making perception systems more intuitive for users by bridging image segmentation with language reasoning, though it is incremental as it builds on existing segmentation and language models.
The authors tackled the problem of enabling image segmentation systems to interpret implicit user intentions by introducing reasoning segmentation, a novel task that uses large language models to reason about instructions and segment target objects, resulting in a competitive performance model and a new 40K dataset for training and evaluation.
Understanding human instructions to identify the target objects is vital for perception systems. In recent years, the advancements of Large Language Models (LLMs) have introduced new possibilities for image segmentation. In this work, we delve into reasoning segmentation, a novel task that enables segmentation system to reason and interpret implicit user intention via large language model reasoning and then segment the corresponding target. Our work on reasoning segmentation contributes on both the methodological design and dataset labeling. For the model, we propose a new framework named LLM-Seg. LLM-Seg effectively connects the current foundational Segmentation Anything Model and the LLM by mask proposals selection. For the dataset, we propose an automatic data generation pipeline and construct a new reasoning segmentation dataset named LLM-Seg40K. Experiments demonstrate that our LLM-Seg exhibits competitive performance compared with existing methods. Furthermore, our proposed pipeline can efficiently produce high-quality reasoning segmentation datasets. The LLM-Seg40K dataset, developed through this pipeline, serves as a new benchmark for training and evaluating various reasoning segmentation approaches. Our code, models and dataset are at https://github.com/wangjunchi/LLMSeg.