Tesseract
PulseAugur coverage of Tesseract — every cluster mentioning Tesseract across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
LLMs struggle with noisy documents, new benchmark reveals
A new research paper benchmarks several open-source large language models (LLMs) for key-value pair extraction from documents, specifically examining their performance under Optical Character Recognition (OCR) noise. Th…
-
Developer builds RAG platform to prevent confident hallucinations
A developer has created RAG.NextUpgrad, a platform designed to prevent retrieval-augmented generation (RAG) systems from confidently hallucinating answers. The platform prioritizes running on low-resource, free-tier hos…
-
MLLMs struggle with low-resource Khmer documents, study finds
A new pilot study has evaluated the capabilities of multimodal large language models (MLLMs) in understanding low-resource Khmer documents. Researchers found that while current MLLMs can process visually clear English a…
-
New dataset and pipeline advance Arabic manuscript OCR research
Researchers have introduced AraMS-28k, a large-scale dataset for historical Arabic manuscript recognition, featuring 28,600 annotated text lines across 14 books. This dataset includes detailed annotations for main text …
-
Pipeline extracts billions of tokens from historical newspapers
Researchers have developed the Institutional Newspapers Pipeline, a modular system designed to extract high-quality, structured data from historical newspaper scans. This pipeline, created in collaboration with the Bost…
-
Tesseract: A robust open-source OCR engine with broad language support
Tesseract is a long-standing open-source Optical Character Recognition (OCR) engine that supports over 100 languages. It is considered a baseline tool for OCR tasks and remains a notable open-source project in the devel…
-
IBM's Docling offers self-hosted PDF-to-Markdown conversion for LLM pipelines
Docling, an open-source document parser developed by IBM, can convert various file types including PDFs, DOCX, and images into clean Markdown or JSON. This tool is particularly beneficial for LLM pipelines as it preserv…
-
Production RAG pipelines require advanced architecture beyond simple demos
This article details the complexities of building a production-ready Retrieval-Augmented Generation (RAG) pipeline, contrasting it with simplified demo versions. It highlights common failure points such as outdated info…
-
Google Research enables quantum computers to learn from errors
Google Research has developed a reinforcement learning framework that enables quantum computers to learn from their errors and self-correct during computations. This approach uses an autonomous agent to continuously adj…
-
Fortran code gains automatic differentiation via LFortran and Enzyme
Researchers have developed a method to enable automatic differentiation for legacy Fortran code, allowing it to be integrated into modern machine learning frameworks like JAX and PyTorch. This approach uses LFortran to …
-
New framework halves security regression in Android malware detection
Researchers have identified and quantified a critical issue in continual learning for Android malware detection, termed "security regression." This phenomenon occurs when malware samples that were previously detected by…
-
LLMs improve reading order reconstruction for historical Armenian newspapers
Researchers have developed a novel method for reconstructing the reading order of historical Armenian newspapers, which present challenges due to complex layouts and limited linguistic resources. Their hybrid approach c…
-
LV-ROVER ensemble boosts Maltese OCR accuracy by 70%
Researchers have developed LV-ROVER, a novel multi-stream Tesseract voting ensemble designed to improve optical character recognition (OCR) for Maltese, a low-resource language. By building a synthetic training pipeline…
-
OCRmyPDF tutorial guides searchable PDF conversion with advanced features
A new tutorial details how to use the Python tool OCRmyPDF to convert scanned documents into searchable PDF/A files. The guide covers advanced features such as sidecar text extraction, batch processing, and optimizing T…
-
Mosaic benchmark suite evaluates differentiable physics solvers
Researchers have introduced Mosaic, a new benchmarking framework designed to evaluate differentiable partial differential equation (PDE) solvers. The framework standardizes gradient access across various solvers, packag…
-
University seeks on-premise document parsing tools for data governance
A university IT department is seeking an on-premise document processing solution to index and search administrative PDFs, class schedules, and meeting notes. Due to data governance policies, cloud-based APIs are not an …
-
Mamba models offer faster OCR but lag Transformer accuracy on historical texts
Researchers have benchmarked State-Space Models (SSMs), specifically Mamba, against Transformers and BiLSTMs for Optical Character Recognition (OCR) on historical newspapers. The studies indicate that while Mamba-based …
-
Baidu's PP-OCRv6 achieves 97ms inference, leads global OCR benchmarks
Baidu's Wenxin officially released the new OCR model PP-OCRv6, offering Tiny, Small, and Medium versions that support over 50 languages and are deployable across various scenarios from browsers to servers. The Tiny mode…
-
AI agent built to safely summarize patient discharge data
This article details the creation of an AI agent designed to summarize patient discharge information from PDF documents. The agent focuses on extracting structured data like diagnoses, medications, and allergies, priori…
-
Baidu's PaddleOCR-VL-1.6 sets new SOTA in document parsing
Baidu's Wenxin has released PaddleOCR-VL-1.6, a new version of its open-source OCR tool. This update achieves over 96.33% accuracy on the OmniDocBench v1.6 benchmark, surpassing major models like Gemini-3-Pro and GPT-5.…