-
Video submission: upload full-body running footage for biomechanical analysis.
-
Athlete identity: manage the profile linked to every performance assessment.
-
Scouting dashboard: view readiness scores, assessment status, and completed performance videos.
-
Performance archive: revisit previous assessments and compare results over time.
-
Explainable results: review readiness scores alongside pose detection and video quality indicators.
-
Gait insights: explore estimated cadence, contact time, symmetry, and stance-versus-swing timelines.
-
Annotated replay: compare original footage with pose overlays and frame-synced joint measurements.
-
Progress analytics: track technique, efficiency, balance, and overall scores across assessments.
-
Athlete login: access your personal sports assessment workspace.
Inspiration
Inspiration Many athletes lack access to specialist coaching and expensive motion analysis equipment. Avishkara aims to make running technique easier to understand by turning ordinary videos into useful movement insights.
How we built it
We built the interface using Next.js, React, TypeScript, and Tailwind CSS. Python and FastAPI power the backend, while MediaPipe detects body landmarks and OpenCV processes and annotates video frames. Movement measurements are estimated using landmark positions, joint geometry, and frame timing. SQLite stores athlete profiles and assessment results.
Challenges we ran into
Lighting, camera angles, motion blur, and partially visible bodies affect pose detection. Identifying gait events from video and converting noisy measurements into understandable feedback were key challenges. We also worked on keeping video processing, authentication, and dashboard updates consistent across the application.
Accomplishments that we're proud of
Avishkara is a sports biomechanics and athlete assessment platform. Users upload running videos to receive annotated footage, joint-angle analysis, estimated gait metrics, and plain-language feedback. The platform examines cadence, foot strikes, toe-offs, ground contact time, overstriding, and left-right symmetry. Athlete profiles, assessment dashboards, readiness scores, and leaderboards help athletes and coaches track and review performance.
What we learned
We learned how to combine computer vision with a complete web application and explain biomechanics to users without technical backgrounds. Building Avishkara also highlighted the importance of recording quality, transparent scoring, and validation against reference measurements.
What's next for Avishkara
We plan to improve gait-event detection, validate measurement accuracy, support more recording conditions, and develop personalized feedback for athletes and coaches.
Log in or sign up for Devpost to join the conversation.