The project is to recognition different traffic signs. It has a two-stage methodology to make the recognition possible, a).Accurate localisation of traffic signs using YOLOv3 tiny model, b). Further classification of detected traffic sign image among various classes using a custom built light weight convolutional neural network. The project is implemented in python, in Google colab and deployed in Nvidia Jetson nano.
Cancellation of noise on speech signal using an adaptive algorithm called least mean square (LMS) algorithm in MATLAB tool.
Region of interest approach applied to ISIC database Melanoma images to segment the cancerous patch. Binary classification implemented using AlexNet, GoogleNet and ResNet. Project implemented in Python