This project intends to summon discrimination indicators in Mexico. The project's background concurs with the 2018 Mexico's presidential and legislative elections and, through statistical analysis, machine learning and computer vision, we are trying to surmise the tendencies concerning skin color for voting. The dataset of public spokesmen and spokeswomen has been acquired already; however, the algorithm for skin detection is flaw.
I need professional support in order to have a flawless skin detection tool that holds up in a python kernel. A complete description of our current status concerning this project is in the html and Jupyter Notebook documents attached inside 'UpWork_SD_Files.zip'.
The key problems to solve are the following:
- Assuring that all images are of the person searched.
- Making a nice skin detection so nothing else gets caught in the process.
- Dealing with makeup.
- Dealing with illumination .
- Develop a precise method to represent the average color of the skin detection with the PERLA palette....
I am looking for a mix of experience and value