LabelU

LabelU 是一款中文的开源数据标注工具。目前具备拉框、多边形、标点、标线、分类、描述等图像标注能力，能够支持目标检测、图像分类、实例分割、文本转写、轮廓线检测、关键点检测等计算机视觉任务场景，通过工具的自由组合即可自定义标注任务，支持COCO、MASK格式数据导出。 欢迎大家在GitHub的Issue里反馈问题和bug！LabelU is a Chinese open-source data annotation tool. Currently, it provides a range of image annotation capabilities including bounding box drawing, polygon annotation, point marking, line annotation, image classification and descriptive annotation. It supports various computer vision task scenarios such as object detection, image classification, instance segmentation, text transcription, contour detection and keypoint detection. Users can customize annotation tasks by flexibly combining these built-in tools, and supports exporting annotation datasets in COCO and MASK formats. Everyone is welcome to report issues and bugs via GitHub Issues!

搜集汇总

背景与挑战

背景概述

LabelU是一款开源的中文数据标注工具，具备拉框、多边形、标点等多种图像标注功能，可支持目标检测、图像分类、实例分割等计算机视觉任务。该工具由上海人工智能实验室于2022年12月16日发布，采用Apache 2.0许可证，并允许用户通过自由组合自定义标注任务，同时支持COCO和MASK格式的数据导出。