Artificial Intelligence
Face Recognition Attendance System
Camera-based attendance with liveness detection and analytics
- Python
- OpenCV
- FaceNet
- Flask
- +2
Artificial Intelligence · Major Project
Fine-grained classification of car make, model and year
Overview
Tells apart cars that look almost identical using the public Stanford Cars dataset of 196 make, model and year classes. The pipeline crops each car with its bounding box, applies heavy augmentation, and fine-tunes pretrained ResNet and EfficientNet backbones. Grad-CAM heatmaps show which parts of the car drove each prediction, and a web app returns the top five guesses for an uploaded photo.
What makes it stand up
Module breakdown
Loads annotations, crops cars and augments images.
Fine-tunes backbones with staged unfreezing.
Top-k accuracy and the most-confused class pairs.
Grad-CAM overlays for correct and wrong predictions.
FastAPI inference behind a React upload page.
After this, you will be able to
Technology stack
You receive
Keep looking
Artificial Intelligence
Camera-based attendance with liveness detection and analytics