Face Match Video Analytics
Face matching for pictures and videos: watchlist, picture search, video search with times, and who appears in two videos. OpenCV YuNet + SFace, offline. Python, Tkinter.
- Price: ₹1,799
- Category: AI / ML
- Technology: Python, OpenCV (YuNet face detector + SFace face recognition), NumPy, Tkinter, Pillow, SQLite, Matplotlib
- Includes: Full source code
- Suitable for: BCA / B.Sc CS, MCA / M.Sc CS, B.E / B.Tech CSE & IT
Project overview
This is a desktop application for Windows with a login and five screens. Watchlist: add people with a name, a note and photos; thumbnails show what is stored (only 128 numbers and a small thumbnail, not the photo). Picture: choose or open a picture; every face gets a green box with the name and the similarity, or a grey Unknown box; the match threshold is yours to set. Video: search a video file every N frames; the faces are shown live, and the result lists each watchlist person with the times they appear, the best similarity, a timeline chart, the number of unknown faces, a CSV file and an HTML report with snapshots. Compare videos: video A is searched for its different people (grouped by similarity, no watchlist needed), each person is looked for in video B, and the result shows who is in both videos and at what times (known people are named from the watchlist). Reports: the log of every search and every sighting with a chart and CSV export. Two demo videos and a demo watchlist of five public-domain portraits are included; in the demo, Armstrong, Glenn and Ride appear in both videos at different times, while two other people appear in only one. 23 code tests and 62 screen checks are included.
Modules
- Watchlist: People with a name, a note and photos; each photo is stored as 128 numbers and a small thumbnail; add, delete, thumbnails.
- Picture search: Finds every face of a picture and names it (or Unknown) with the similarity; threshold setting; save the result picture.
- Video search: Checks every N-th frame, shows the faces live, lists when each person appears (appearances merged), timeline chart, CSV and HTML report with snapshots, stop button.
- Compare videos: Groups the faces of video A into different people and looks for each in video B with the times; names known people; CSV and HTML report.
- Face engine: OpenCV YuNet detects faces, SFace makes the 128 numbers; cosine similarity; greedy grouping of the same person; merging of sightings.
- Reports and log: Log of searches and sightings in SQLite with a chart and CSV export.
- Login: Admin login with a salted hash and a change-password dialog.
Requirements
- Python 3.11
- Internet once, to install the libraries (about 3 minutes)
- Windows 10 / 11
How to run this project
- Install Python 3.11 from python.org and tick "Add python.exe to PATH" during setup.
- Extract the zip and open the project folder (STUDENT_GUIDE.md explains every step in simple words).
- Double-click `1_INSTALL.bat` once and wait until it says [OK] (about 3 minutes; it downloads the libraries into a private .venv folder).
- Double-click `2_RUN.bat` to start the project. If it does not start, double-click `CHECK_SETUP.bat` and read what it says.
- Log in with admin / admin123.
- Compare videos: press 'Who from video A is also in video B?': five different people in each video, three seen in both with their times in B; export the CSV. Reports: show the log of searches and sightings and the chart.