PopularMajor Project
Artificial Intelligence
Face Recognition Attendance System
Camera-based attendance with liveness detection and analytics
- Python
- OpenCV
- FaceNet
- Flask
- +2
₹4,900₹5,999
4.8/301 reviews
Artificial Intelligence · Capstone
Crowd Counting & Density Estimation
Density-map regression that counts people in packed crowds
4-5 weeks build
Overview
When a crowd is too dense for bounding boxes, this kit regresses a density map instead. A CSRNet-style network with a VGG front end and dilated convolutions is trained on the ShanghaiTech dataset, and summing the predicted map gives the head count. A heatmap viewer highlights overcrowded regions and raises a warning when a set capacity is crossed.
What makes it stand up
- 01Ground-truth density maps generated with adaptive Gaussian kernels
- 02CSRNet-style dilated convolutional regressor
- 03Trained and tested on the ShanghaiTech Part A and B splits
- 04Heatmap overlay with capacity warnings per region
- 05MAE and MSE reported against a detection-based baseline
Module breakdown
- 01
Dataset prep
Head annotations converted into density maps.
- 02
Model
VGG front end with a dilated back end.
- 03
Training
Patch sampling, loss curves and checkpoints.
- 04
Inference
Count, heatmap and region-wise density on new images.
- 05
Viewer
Streamlit app with capacity threshold alerts.
After this, you will be able to
- Explain why density maps beat detection for dense crowds
- Defend your kernel choice for ground-truth maps
- Demo counting on a festival or station image
Technology stack
- Python
- PyTorch
- OpenCV
- NumPy
- Streamlit
- Level
- Capstone
- Domain
- Artificial Intelligence
- Build time
- 4-5 weeks
- Difficulty
You receive
- Complete, commented source code
- 60-90 page project report (IEEE format)
- Editable presentation deck
- Architecture, ER and UML diagrams
- Local setup and deployment guide
- Sample dataset or seed data
- Viva question bank with answers
- One 45-minute walkthrough call
- 30 days of doubt-clearing support