Enhances Tesseract OCR output using LLMs (local or API) for error correction, smart chunking, and markdown formatting of scanned PDFs

Using yolov4 object detection and image processing with OpenCV, the project enables detecting and reading licence plates and storing a log plates detected. The detection is achieved by detecing all the vehicles in the photo using a yolov4 tiny model that was converted to tflite and for each of the detected vehicles using image processing the lince plate is detected and after enhancing the contrast between the letters and the background the letters the licence plate is being processed by tesseract-ocr and we get the licence plate number.

This research aims to fine-tune an Arabic OCR model using Tesseract 5.0, enhancing text recognition accuracy through extensive data collection, preprocessing, and image generation. By leveraging advanced training techniques and data augmentation, we achieve significant improvements in word error rates (WER).

Preprocessing methods to enhance Tesseract-OCR in the case of printed text on difficult background, or handwritten text on lined/squared paper.

The aim of project is the development of an automated text recognition system with the support of the Ukrainian language, which focuses on the pre-processing of images and processing of recognized text. Two models of machine learning have been developed: to determine the angle at which the image is rotated and to determine the type of document. Algorithms for image enhancement (binarization, noise removal, contrast adjustment) and an algorithm for correcting errors in text using fuzzy string logic and Levenshtein distance have also been developed. The system is presented as a desktop application. The server part was developed using the C# programming language and ASP.NET Core, ML.NET frameworks, client part - WPF (Windows Presentation Foundation), text recognition - Tesseract OCR.

GPU-enhanced OCR pipeline for historical newspapers using Google Vision or Tesseract

This project is a modified version of the charlesw/tesseract library, designed to enhance OCR functionalities within the .NET ecosystem by improving cross-platform support. It aims to make the Tesseract OCR engine more effective in .NET environments.

An image processing project that captures vehicle images and extracts the license plate using OpenCV, enhancing the clarity of the plate using image enhancement techniques. Tech Stack: Python, OpenCV, NumPy, Tesseract OCR Skills Highlighted: Computer Vision, Python, Image Processing

Pharmacist Assistant - AI-Powered Prescription Reader An AI-driven tool that extracts handwritten prescription text using OCR (Tesseract) and validates medicine names. Built with Flask, OpenCV, and AI, it enhances accuracy, reduces errors, and streamlines pharmacy workflows by automating prescription processing

Enhancing Image quality for better understanding by Tesseract OCR

A lightweight, end-to-end document scanner built using OpenCV and Tesseract OCR. This project detects documents (like receipts, paper, forms), extracts the perspective-warped version, enhances it, and finally uses OCR to extract the textual content from the image.

Developed an OCR-based ANPR system using Python, OpenCV, and Tesseract, achieving 90% text extraction accuracy. Integrated a SQL database for real-time license plate data storage and retrieval, enhancing traffic monitoring.

This project automates text extraction from Hindi/Sanskrit PDFs using a pipeline that converts pages to images with pdf2image and applies OCR via Tesseract. It cleans and segments text based on Devanagari punctuation, preserving key elements like dates. The output, structured for NLP tasks, enhances accessibility to Indian language documents.

Android Native app that demonstrates a heavy using of NDK (C++) for performing advanced image preprocessing (Sauvola Thresholding / Binarization) for enhancing Tesseract OCR engine results.

In this project, I created an OCR post-processing procedure to enhance Tesseract OCR output

Bank Cheque OCR automates cheque data extraction using OpenCV, PIL, and Tesseract OCR, enhancing accuracy in retrieving key details like cheque number, account number, date, and amount.

Team HackAttack:Our solution combines state-of-the-art technologies to enhance web accessibility for visually impaired users. Leveraging the Vision Transformer model, we achieve real-time image recognition, accurately identifying objects and text within images. The integration of the Tesseract OCR API enables dynamic text extraction from images, pr

This Python code leverages the EAST text detection model and Tesseract OCR to detect and recognize text in images, enhancing text extraction for various applications.

Enterprise Content Management MVP with semantic search capabilities. Upload PDFs and images OCR text extraction from images using Tesseract LLM-assisted metadata suggestion using Google Gemini (enhanced with OCR text) Semantic and exact metadata search Image similarity search using OpenCLIP Vector embeddings with Qdrant SQLite database

A Python-based OCR tool for extracting text from PDF and image files using Tesseract OCR. The script includes advanced image preprocessing techniques to enhance text recognition accuracy and outputs extracted text as .txt files. Ideal for processing scanned or handwritten documents.

Al-powered document scanning and processing system with real-time desktop-mobile synchronization. Built with Flask (Python) backend, React + TypeScript frontend, OpenCV image enhancement, Tesseract OCR and Socket.IO WebSockets for seamless printing and workflow management.

Made enhancements in the process of detection and recognition of license plate images with YOLOv8 algorithm in an object detection system and the classified images sent through CNN and Tesseract OCR based combined algorithms for the character recognition process.

wins_cc is a lightweight, real-time English-to-Chinese interpreting tool designed for Windows 11, built with Python. It enhances the native Live Caption feature by adding automatic OCR (via Tesseract) and seamless machine translation (via Chrome’s built-in translation API).

Developed an innovative real-time Image Text to Speech (ITTS) system using Python, OpenCV, Tesseract OCR, and Google Text-to-Speech API to enhance accessibility for visually impaired individuals. The system accurately extracts text from live video streams and converts it into natural speech.

AI-powered medical report analyzer that extracts text from PDFs/images, summarizes reports, detects abnormalities, and provides a chatbot for medical queries. Built with FastAPI, OCR (Tesseract, pdfplumber), OpenAI GPT-3.5, and deployed on Google Cloud. Future enhancements include medical image classification and predictions. Contributions Welcome!

This project is a YOLOv8n-based license plate recognition system using transfer learning and CUDA for fast detection. OpenCV enhances plate visibility, Tesseract OCR extracts alphanumeric characters, and text is cleaned for accuracy. A dark-themed Tkinter interface enables easy model loading, image selection, and result display.

A powerful web-based OCR application built with Flask, supporting both Tesseract and EasyOCR engines. It provides high-accuracy text extraction with multi-language support, GPU acceleration via PyTorch and CUDA, and image preprocessing using CLAHE to enhance recognition. User-friendly interface for real-time image-to-text conversion.

This Python project enables the extraction of data from scanned PDFs of voter lists in Hindi, sourced from the official government electoral roll website. Using Tesseract OCR and PDF enhancement tools, it converts scanned PDFs into structured Excel files with columns for Name, Father/Husband Name, House Number, Age, and Gender.

Developed and enhanced .NET-based components for document management systems, including OCR integration for Persian documents, Enahanced Persian Data Files include more efficiency according to Default tesseract files.

A simple OCR tool that converts Sinhala PDFs and images into editable Word documents using enhanced Tesseract OCR