2026-10-13T01:50:24Z https://www.scidb.cn/oai

10.57760/sciencedb.013fj 2026-09-28T11:32:54Z

2026-09-28 A Dataset for Scene Text Referring Expression Segmentation doi:10.57760/sciencedb.013fj en TextSeg is a dataset for scene text segmentation, designed to support pixel-level localization of text regions in natural and designed images. The dataset contains scene images and detailed text annotations, including text region masks and character-level annotations.BTS (Bi-lingual Benchmark for Text Segmentation in the Wild) is a bilingual benchmark dataset designed for scene text segmentation in natural images. Scene Text Segmentation ; Computer Vision; Image Segmentation; Scene Text Referring Expression Segmentation 李增 RESTRICTED dataset Science Data Bank