With a minimal codebase in place, I began experimenting with turning the body segmentation mask into a blob. After exploring the TensorFlow libraries’ documentation, I was pleased to find a command that could extract the contour, or outline, of the segmentation. With the body’s outline in hand, I needed a way to manipulate and process it.
Naturally, I turned to OpenCV—a library I’ve worked with in Python but never in Node.js. Configuring OpenCV for Node
By March 13, the graduate thesis demo day, I hope to replace the current polygon-shaped prototype with a fully amorphous and organic silhouette generated from the webcam feed. To achieve this, I'll likely need to configure OpenCV for Node.js at some point in the next four weeks. I plan to focus solely on the "blobification" aspect of the installation, leaving the networking and generative components for later since they will require more time and I am less familiar with them. Ideally, I aim to have the webcam feed "blobification" mostly completed within the next two weeks so that I can get a headstart on the networking side before the demo even rolls around.