A deep learning project that leverages U-Net architecture and transfer learning with pretrained CNN models to restore blurred text images. Built with TensorFlow/Keras and includes a Streamlit web application for easy demonstration.

Developed an AI-powered image processing tool that uses a Deep learning colorization model (Caffe) with 86% accuracy to restore colors in black-and-white images. Implemented a Wiener filter-based deblurring algorithm with Fourier transforms to enhance image clarity and used PyTesseract and OCR to extract text from image.

> 🤖 David One AI is a fully offline multilingual AI assistant that supports voice and text input, image editing, enhancement, deblurring, and intelligent coding. Built using the Arya Framework, it works after one-time setup without internet. Supports 9+ Indian languages including Hindi, Bengali, Tamil, and more