Data Analyst
Video
Object DetectionSCOPE Project objective- To annotate a specialized video dataset for training, validating, and testing computer vision models. input data profile- Raw video footage in standard formats(MP4,AVI) captured at specific resolutions(1080p, 4K) and frame sets(30fps, 60fps). Target environment- Specific operational conditions captured in the video such as lighting, weather, geographical locations and camera angles. SPECIFIC DATA LABELLING TASKS Temporal video segmentation- Cutting long video files into short, meaningful clips based on specific events. Skeletal Keypoint Annotation- Mapping specific anatomical or structural coordinates on objects to track posture, movement, and orientation. QUALITY MEASURES ADHERED TO Annotation Guideline Alignment- Conducting training and testing for all annotators. Automated Validation Scripts- Running programmatic syntax checks to catch overlapping boxes, missing labels. Error correction loop- Flagging rejected annotations.