Deep Learning Projects

Deep Learning Systems: Super Resolution GAN | Python

• Implemented Single Image Super-Resolution (SISR) using SRGANs, incorporating transfer learning, depthwise convolution, and attention layers. The model was evaluated using perceptual loss, adversarial loss, PSNR, and SSIM metrics on celebHQ and div2K datasets to enhance image quality.

ASL Classification for 3D Point Cloud Using PointNet | Python

• Developed a system to classify American Sign Language (ASL) gestures using 3D point cloud data captured by a Pointflex camera. Utilized the PointNet architecture for effective preprocessing and classification, facilitating a real-time communication environment for ASL users.

Deep Learning Concepts

• MNIST Claasification
• Speech Denoising
• Network Compression
• Generative Models