Mask Detect play online

The purpose of this application is to demonstrate the viability of an image recognition system that allows detecting if a mask is correctly fitted. You just stand in front of the camera and the application will inform you if you have the mask on correctly. Otherwise, it will inform you of the possible problem through a voice message.

The training set has been created thanks to citizen participation. Thousands of volunteers have uploaded selfies with different mask placements, using the application "Intemasc Captura" https://play.google.com/store/apps/details?id=es.upv.mastermoviles.intemasc.captura. Once the training set was compiled, the nursing group of the General Hospital d'Ontinyent (FISABIO) has labelled each type of problem.
Although the type of problems studied is wide, in this demo only those that have worked best are collected: No mask, Bad adjustment of nasal bridge, Nose out and Correct.

The Deep Learning models are trained using the following code: link
The following article describes the development of the project: link

PRIVACY POLICY: The data processing is carried out in real time. No images are stored or data is sent over the internet.