README.md



Show me the color

This is Personal color diagnosis system.


Click this image to see the Demo video!


Usage

cd src


Analysis a single image

python main.py --image IMAGEPATH

Analysis multiple images in a directory

python main.py --dir DIRECTORYPATH


1. Face detection

detect_face.py with shape_predictor_68_face_landmarks.dat has DetectFace class, and it provides face detection function, the exact face parts, and the coordinates of them. We selected cheeks, eyes, eyebrows(instead of hair) for personal color analysis.


2. Extract the Dominant Color

dominant_colors.py has DominantColors class, and it provides the dominant colors by k-means algorithm, with RGB values.
I modified the source code from (https://buzzrobot.com/dominant-colors-in-an-image-using-k-means-clustering-3c7af4622036).


3. Personal Color Diagnosis

The RGB values from step 2 is converted to Lab and HSV color space. The b value from Lab is the factor determining warm/cool and the S value from HSV is the factor determining spring/fall or summer/winter. Our team got the criteria values which classifies the personal color results by analyzing color values dataset from several images.

tone_analysis.py and temp5.py is the source code for personal color classifying and the usage.


4. Virtual Makeup Simulator

We classified several lipstics as 4 personal colors by their colors and put their name, color code and class into the Database. After detecting the lip outlines, the system connects the lips and puts the chosen color from the DB.


5. Web

Django is used for web framework.