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
Lip Reading with Neural Networks
Visual speech recognition that transcribes sentences from mouth video
5-6 weeks build
Overview
Reads sentences from silent video. Faces are detected, lips are cropped from facial landmarks, and the mouth-region sequence goes through spatio-temporal convolutions and bidirectional GRUs trained with CTC loss, so no frame-level alignment is needed. Trained and evaluated on the GRID audio-visual corpus, with word and character error rate in the report.
What makes it stand up
- 01Mouth-region cropping and alignment from facial landmarks
- 023D convolution front end with bidirectional GRU layers
- 03CTC loss so training needs only sentence-level transcripts
- 04Greedy and beam-search decoding compared
- 05WER and CER on held-out speakers in the report
Module breakdown
- 01
Data preparation
GRID video loading, landmark detection and mouth crops.
- 02
Model
Spatio-temporal CNN into stacked bidirectional GRUs.
- 03
Training
CTC loss, augmentation and checkpointed training runs.
- 04
Decoding
Greedy and beam-search decoders over the character set.
- 05
Demo
Streamlit app that plays a clip beside its predicted transcript.
After this, you will be able to
- Explain how CTC removes the need for frame-level labels
- Defend your speaker split and why it matters for evaluation
- Demo transcription on held-out clips and discuss the errors
Technology stack
- Python
- TensorFlow
- OpenCV
- dlib
- Streamlit
- Level
- Capstone
- Domain
- Artificial Intelligence
- Build time
- 5-6 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