LabelLab - Crowdsourcing and Data Collection
Published in Both Android and iOS, 2023
LabelLab is an inventive application that marries the engaging user experience of TikTok with the practical functionality of a data labeling tool. This unique platform is designed with dual objectives: to provide a social media platform where users can watch, share, and enjoy video content, and to harness the collective power of its user base for the purpose of improving artificial intelligence research in emotion and human activity recognition.
At its core, LabelLab allows users to actively participate in the creation of a comprehensive video dataset by labeling the content of videos as well as the hashtags associated with them. This process is not limited to the content users upload themselves; it extends to all videos and hashtags within the app, fostering a community-driven approach to data annotation.
What sets LabelLab apart is its capability to function as a crowdsourcing labeling app. This feature enables the importation of external video datasets into the platform, thereby broadening the scope of data available for labeling. To facilitate efficient data annotation, LabelLab provides tools for recruiting and managing crowdworkers. These crowdworkers, drawn from the app’s user base, are instrumental in labeling video data, making the process faster, more efficient, and scalable.
You can find the code for this app in this public repository on Github.