Background:
Using facial expressions is an effective way to convey emotions for which we can get reciprocal feedback during conversations. However, the world we live in today has changed so much that those spectacularly vivid expressions are hidden behind a layer of the face mask; we have come to rely on the sound of speech to judge the emotions of others, but this is not always accurate, especially when we are dealing with strangers, and it becomes more difficult to feel the emotions of others. I don't know how long this disconnect will last, not to mention that for some specific groups of people, the situation of wearing a mask can last decades or even a lifetime; therefore, I desperately hope that this project will provide an effective solution to the communication disconnect.
Concept Statement:
I am creating an inclusive mask solution to help people communicate and transmit expressions. It is an accessory that can be fitted to a regular mask, breaking down the barriers brought by face masks to interpersonal communications through speech and facial expressions.
Implementation:
I want to divide my implementation process into two parts, one is detecting the wearer’s expression and the other is actually showing the detected expression. When we make certain expressions, the position and strength of the contact between our face and the mask changes and that’s the basic methodology of the detection process. I will put several conductive materials on the mask and each of them connects to the touch sensing pins of the ESP32 board. So when I make an expression with the mask on, I will get a set of values and different expressions will have their independant data sets. And when I collect enough data sets, I can import the csv files into the machine learning, training and classifying the data. By doing this, the detection accuracy will be higher. And for the visualing part, I want to use a small flexible OLED screen that can be put on the mask. It’s very compact and can be bent a little in order to fit the mask shape.
Early Prototypes:
For the very first prototype, I just sewed some conductive threads on the mask; and I can use alligator clips connecting each of the thread to the capacitive touch sense pin on ESP32.
Quick demo videos:
After the quick test, I found out that using alligator clips was good for fast testing, while they would influnce reading when they were touched and moved. Also, using conductive thread was not a comfortable solution for the wearer so I might need to switch to other conductive materials.