Initiated by the University of Michigan Herbarium, VoucherVision harnesses the power of large language models (LLMs) to transform the transcription process of natural history specimen labels.
Instead of using LM2 to crop out labels, add tool to create label collage from already cropped images. This will parse file names and create a "specimen" image from the parts.
e.g. FMNH_809_a.jpg + FMNH_809_b.jpg + FMNH_809_c.jpg --> FMNH_809_collage.jpg
Instead of using LM2 to crop out labels, add tool to create label collage from already cropped images. This will parse file names and create a "specimen" image from the parts.
e.g. FMNH_809_a.jpg + FMNH_809_b.jpg + FMNH_809_c.jpg --> FMNH_809_collage.jpg