22 lines
926 B
Markdown
22 lines
926 B
Markdown
## Project Description
|
|
|
|
Javascript Canvas face recognition using back buffer for performance optimization.
|
|
|
|
Mixing convolution, substractive filter, this project tends to localise the components of the human face in the first hand, and in a second hand, to use ratios to identify faces.
|
|
|
|
The final aim is to be able to detect face components in **real-time** to control a program with your face only. Also, it must recognize components for whatever picture/video size, whatever fashion accessory weared.
|
|
|
|
|
|
## Development timeline
|
|
I. Image optimization
|
|
1. Size adjustment
|
|
2. Contrast adjustment
|
|
3. Brightness adjustment
|
|
4. Edge detection
|
|
II. Face extrusion
|
|
1. Set ratio parameters (human face ratios and constants)
|
|
2. Set face limits
|
|
3. Remove background
|
|
III. Components recognition
|
|
1. Set ratio parameters (human face ratios and constants)
|
|
2. Set substractive filters (soft deep learning) |