Jhanvi Patel

License Plate Detection: Implemented license plate detection using YOLOv5 in PyTorch for both Indian and foreign number plates, achieving a remarkable 95 percent accuracy. Deployed the production pipeline on Raspberry Pi. Motion Detection: Crafted a robust surveillance system utilizing machine learning techniques such as Background Subtraction for precise motion detection. Keypoint Regression model: Developed a keypoint detection model leveraging the O-Net architecture to accurately map and align license plates.

Driver Monitoring System: Implemented a Driver Monitoring System utilizing the SSD model architecture to classify four distinct driver activities: smoking, yawning, using phone, and eating. Successfully deployed the model as an Android TensorFlow Lite (TFLite) model. Pose-Estimation for Physiotherapy App: Integrated MediaPipe framework for accurate pose estimation, ensuring precise tracking of users’ movements tracking keypoints for 15 exercises. Acne Classification: Developed State-of-Art model using Convolution Neural Network from Scratch for Acne classification and attained 89 accuracy.