basic_makeup

This repository is intented for the implementation of automated python scripts for applying basic makeup to images using computer vision tools.

Description

This repository contains the implementation of the following makeup "filters":

1. im_lipstick
2. im_eyecolor

The scripts have been implemented using python, as part of the course Computer Vision II - Applications at opencv. The example images used in this repository are taken from the lfw dataset and one image was provided by the course.

Both scripts display four outputs: the input image (original size), the gray-mask (scaled), the "colored" mask (scaled) and the output image (scaled) after applying the corresponding filter. However, the resulting image is saved to the same size as the input image.

Pre-requisites

Before running the scripts, please make sure to install in advance the following:

• python version 3.8
• opencv
• dlib for machine learning

Usage of automated lipstick

Navigate to the script directory and type in a terminal: python im_lipstick.py -f path_to_file.jpg -a hue_value --saveIt

Explanation:

• -f refers to the full path of the image you want to apply lipstick.
• -a refers to the hue value. This parameter controls the color of the lipstick.
• --saveIt implies to save the image with the applied lipstick. Please omit this if you do not want to save the result.

This script provides a very basic way to apply lipstick to a given image. It uses the HSV color model to do that. If you are familiar with the RGB color model, then you can think of HSV as the "cylindrical coordinate system" for color, being RGB, the "cartesian coordinate system". Hue gives the "sense" of color in human vision. Further explanation about the HSV color model can be found in the Wikipedia and a good resource to find the actual value of hue for a desired lipstick color can be found here.

One of the main limitations of this implementation is that it assumes the lips are closed in the input image. Therefore, it will fail to "nicely" apply the lipstick if the lips are wide open in the given image. Also, only the hue is used as input parameter for chosing color.

Usage of automated eye-color

Navigate to the script directory and type in a terminal: python im_eyecolor.py -f path_to_file.jpg -a hue left-iris -b hue right-iris --saveIt

Explanation:

• -f refers to the full path of the image you want to change the eyes color.
• -a refers to the hue value of the left iris of the face in the image. This parameter controls the color (see above).
• -b refers to the hue value of the right iris of the face in the image. This parameter controls the color (see above).
• --saveIt implies to save the image with the applied lipstick. Please omit this if you do not want to save the result.

In this case, the hue of each iris can be set independently (as in the example below). The main limitations of this implementation is that only the hue is the control parameter, therefore not all the aspects of the color-space are covered.

Sample results

Input image	Lipstick applied
Input image	Eye-color applied

TODOs

• Implement the case when the lips are not close to each other in the lipstick script.
• Implement additional filters: blush, hair color,...