4.1.1 Types of Data Quests
FileMarket AI provides diverse data collection tasks to enhance AI model training. All quests involve self-labeling, and future improvements will include AI-assisted validation and annotation double-checking for greater accuracy.
Image & Gesture Quests
Users submit selfies, full-body photos, palm images, and gesture-based actions under varied conditions (lighting, angles, backgrounds) to support computer vision, identity verification, and biometric AI models.
Audio Quests
Participants record speech samples, dialects, and accents for speech recognition, text-to-speech (TTS), and voice AI models. Some tasks focus on regional accents or scripted prompts, while others require natural conversations.
Video Quests
Users provide short video clips demonstrating facial expressions, gestures, motion, and object interactions, improving AI-driven motion tracking, behavioral analysis, and AR/VR applications.
Labeling & Annotation Quests
Participants verify and refine AI-generated labels on images, audio, and video data to enhance dataset accuracy. Future tasks will involve AI-assisted pre-labeling, with contributors reviewing and correcting annotations.
Specialized Data Quests
High-value datasets tailored to custom AI needs, including:
Conversational Datasets (dialogues, speech interactions)
Environmental Sound Captures (background noise for AI recognition)
Handwritten Text Data (OCR training)
Custom AI Training Sets (specific requests from AI companies)
Last updated