AutoImageSeg: A zero-code image segmentation software toolkit
Weihao Gao, Jiarou Lu · SoftwareX · 2026
AutoImageSeg is a zero-code, open-source image segmentation software toolkit that integrates nine mainstream models. It offers a closed-loop workflow encompassing training, inference, evaluation, and re-annotation. Through its graphical user interface (GUI), users can effortlessly benchmark models, predict new data, and auto-generate editable LabelMe labels—all without any programming. This streamlined process facilitates rapid iteration and high-quality ground-truth accumulation, especially in small-sample scenarios. By accelerating dataset construction across multiple domains, AutoImageSeg serves as a powerful tool for both researchers and industry professionals.