This article is automatically generated by AI. There may be omissions or inaccurate descriptions in the content. Information sources are X posts, GitHub updates, Current Awareness Portal.

This is a weekly summary of new tool development and release information related to Digital Humanities (DH).

Expanded Character Recognition and Speed Improvements Implemented in Collaborative Transcription Kuzushiji OCR

@yuta1984 has expanded the number of recognizable characters in the Collaborative Transcription Kuzushiji OCR from 5,000 to 7,700 characters. The recognition accuracy has not decreased according to the report.

Additionally, significant speed improvements have been implemented. In WebGPU-compatible environments, recognition time has been reduced from approximately 90 seconds to about 10 seconds. Furthermore, sequential display of recognition results has also been supported.

LLM-Based Modern Translation and IIIF Image Loading Functions Added to Collaborative Transcription Kuzushiji OCR

@yuta1984 has added LLM-based modern translation functionality and IIIF manifest image loading capability to the Collaborative Transcription Kuzushiji OCR. This feature is based on an idea by @tpWacky.

The LLM can be selected from Claude, ChatGPT, and Gemini models, and can be used by users obtaining their own API keys and configuring them in the application.