Data Annotator
Image
PolygonThis project focuses on annotating visible barrier, wall, and fence structures using polygon segmentation. Annotators must outline only the visible portions of the primary barrier in each image and assign the correct material‑type label. The goal is to train models to recognize different boundary structures, understand their shapes, and distinguish material categories such as concrete, metal, wood, and temporary barriers. Annotations follow strict rules for occlusion, continuity, and polygon quality. When exported to COCO JSON, each polygon becomes a segmentation entry with its corresponding category ID, bounding box, and area.
2023 - Present