Computer Vision API V2 - FastAPI & ONNX Models

This repository features two OCR projects: V1 is a Python-based interface integrating Tesseract for local text extraction and the Gemini API for advanced refinement. V2 focuses on deep learning, featuring a custom CNN model trained on the A-Z dataset to recognize handwritten characters. A complete university-level computer vision portfolio.

video-to-tree-species is a computer vision + geospatial pipeline that processes video footage to: 🎯 Detect trees using Grounding DINO (zero-shot object detection) ✂️ Segment trees using SAM2 (Segment Anything Model v2) 🌱 Classify trees into species using the PlantNet API 📦 Save a structured dataset of labeled tree images