data labels
the focus of the project was structured video data labelling to support computer vision model training, the objective was to accurately annotate real-world human activities and object interactions to improve activity recognition and behavioral classification systems. The specific labeling tasks included labelling short form and mid-length videos depicting general activities such as cleaning ,repairs, identifying individual atomic actions relevant to the video tasks across various real-life activity categories and labelling correctly, ensuring the labelled actions corresponded to the timestamps accurately, identifying objects and tools used, environmental context e.t.c. while adhering to quality measures by strictly following annotation guidelines and taxonomy standards, performing self review before submissions to reduce labelling errors, maintaining consistency and high accuracy across repeated action types and adhering to data confidentiality and project compliance requirements.