rmtheis/tess-twoArchived
Tess Two
Tess-two is an optical character recognition tool and Android application designed to extract written text from images using the Tesseract engine. It functions as an image analysis utility for detecting visual artifacts, blur, and optical flow within local image files on Android devices.
The project includes an image pre-processing suite used to clean and manipulate images to increase the accuracy of text recognition. This involves a pipeline that applies grayscale conversion and binarization before the recognition process.
The software integrates native image processing and character analysis to perform on-device visual analysis. It utilizes the Android Native Development Kit and Java Native Interface to handle raw bitmap processing and native code execution.
Features
- OCR Engines - Integrates the Tesseract OCR engine to convert image pixels into machine-readable digital text.
- Image Text Extractions - Extracts written text and characters from images using an integrated optical character recognition system.
- Optical Character Recognition - Implements technology for converting images of text into machine-encoded text.
- Mobile OCR Integrations - Provides a mobile OCR integration that allows Android applications to perform local text extraction.
- Android Implementations - Provides a dedicated Android implementation for extracting written text from images.
- Image Processing - Implements pixel-level image transformations to clean and manipulate images prior to the OCR process.
- Image Pre-processing Utilities - Ships a suite of pre-processing utilities for binarization and noise reduction to improve OCR quality.
- Visual Artifact Detectors - Provides detectors for identifying visual artifacts, blur, and optical flow within local image files.
- OCR Applications - Provides a full Android application for extracting text from images using the Tesseract OCR engine.
- Image Processing Pipelines - Ships an image processing pipeline that applies grayscale conversion and binarization to optimize images for text recognition.
- On-Device Image Recognizers - Performs visual character and artifact recognition entirely on local Android hardware.
- Image Pre-processing - Includes utilities for cleaning and correcting images on-device to improve the accuracy of subsequent OCR.
- Android Analysis Tools - Functions as an Android-based utility for detecting blur, optical flow, and visual artifacts in images.
- Java Native Interface Wrappers - Implements Java Native Interface wrappers to expose high-performance C++ image processing binaries to the Android Java layer.
- Native-Layer Media Processing - Processes raw image bitmap data directly within the native C++ layer to maximize processing speed and minimize bridge overhead.
- NDK Native Implementations - Utilizes the Android Native Development Kit to implement high-performance visual artifact detection in native code.
