人工智能图像识别项目
Image
ClassificationProject Overview Participated in a computer vision model training data construction project, responsible for image classification and object detection data annotation, and also involved in image annotation work related to content review. Served as a Data Annotator, dedicating approximately 20 hours per week, and processing about 50,000 images in total. Annotation Tasks - Use LabelImg tool to complete image classification label annotation and object detection bounding box selection - Judge the compliance of images and label violations according to review rules - Perform self-inspection of data, mark unclear samples for confirmation, and avoid subjective mislabeling Quality Control Annotation accuracy target ≥95%, strictly follow project annotation specification documents; the team implements a sampling inspection mechanism, regularly participates in annotation rule review and correction training to ensure data consistency and delivery quality.