This week I researched different neural networks and architectures we could incorporate (RNNs, CNN) and read up on tensorflow and keras to familiarize myself with the actual programming side of the classification stage. I’m still working on the feature recognition and extraction stage but for the time being, I will just try reducing the dimensions of the processed image and removing the background, and pass in the edge detected image for the static sign recognition. The complexity and time to run this will be the benchmark to beat for different feature extraction strategies we try out.
Carnegie Mellon ECE Capstone, Fall 2020. Team B6; Aaron Selesi, Young Woo Kim, Malcolm Fitts