Video Analytics : Built a video analytics system using Nvidia's deepstream which can handle multiple CCTV cameras and run models like person detection, mask detection, entry and exit counts.Lip Sync : Trained a Conditional Gan network to generate lips in sync with given audio for a target speaker. Further Super Resolution was used to improve the quality of the lips.Worked on multi label classification, to classify palm fruits into different categories like grade and stalk. On the limited dataset got an accuracy of around 75 - 80%.

Worked on sports visual tracking and action identification mainly for badmintonGenerated heatmaps for player movement on the court and identified shotsand strokes using LSTM.Collaborated with Star Sports for live match analysis in Premier BadmintonLeague (PBL) 2020.

Designed an efficient brand logo detection model using YOLO algorithm that could detect various brands appearing on Youtube videos.Also worked on project based on ad targeting which consists of classifying Youtube videos as adult or safe for playing ads on the basis of thumbnails using neural networks like VGG 16 and Resnet-50.

Along with three other members developed a Handwriting recognition model using Deep Learning and deployed it successfully on the android platform using Tensorflow.