The title of my project

Detection of American Sign Language Alphabets

Please watch the below video for a short demo

Demo with two people

Demo with three people

Detection of American Sign Language (ASL) alphabets using the You look only once object detection algorithm (YOLOV4)

Data set collection and Process: I collected around 100-200 images for every alphabet on my own and have done the data augmentation to get around 400 images for each alphabet, which resulted in about 20000 images for the whole dataset. All the images are then labeled using the labelImg tool and trained the YOLOV4 model by keeping max batches to 70000 in the config file on google colab GPU.

Result: Above machine learning model detects the ASL alphabets and shows the predicted alphabet with confidence