GSoC 2026: OpenVINO-Integrated YOLO Video Analytics Pipeline, Wanting Alignment Feedback

[tani-dubey]

Hi OpenVINO team 👋

I’m a Computer Science undergraduate preparing for GSoC 2026 and wanted to share an OpenVINO-integrated prototype I’ve been working on, and get feedback on alignment and next steps.

Project overview:

• End-to-end video analytics pipeline for vehicle & license plate recognition
• Vehicle detection using YOLOv8
• SORT-based tracking
• License plate detection + OCR
• CSV / JSON event logging
• FastAPI-based inference & metrics endpoints

Vehicle detection stage supports OpenVINO inference on CPU

Repo: https://github.com/tanii1125/openvino-vehicle-analytics.git

What I’d appreciate feedback on:

• Does an OpenVINO-integrated video analytics pipeline like this fit the kind of work OpenVINO mentors look for in GSoC?
• Would refining this into a reusable OpenVINO sample / reference pipeline be a meaningful GSoC-style contribution?
• Are there particular areas where you’d recommend focusing to better align with GSoC 2026 expectations?

I’m aiming to refine this into a GSoC-ready contribution and would value any guidance.
Thanks for your time!