I've thought about this for a long time. I tried to follow the process to make image paths, and ground truth text files, and all failed. I was never able to train a model to recognize some of my images. I don't think I've ever seen a tutorial where this is done in a clean manner. Like a 1-2-3-4...-10 steps and you should be able to see XYZ result.
If you know of such a tutorial or even some links that may help uncover the mystery, please share. Thank you kindly. Looking forward to your feedback and hearing from you. PS: if you ever decide on doing this as a GUI, I'd be interested in collaborating, and helping out achieve this GUI. On Thursday, 7 September 2023 at 05:35:27 UTC+1 [email protected] wrote: > I had the same problem recently. I've been thinking about developing a > simple GUI that makes it easy to fine-tune a model. Should be able to just > upload a CSV containing image paths and ground truth text, choose the PSM > value, decide which model to base the training on, and click to start the > training process. Could handle eval on validation set too. Would this be > useful for you? Just let me know if there are other features you'd like to > see for your use case. > > On Sunday, September 3, 2023 at 5:12:05 AM UTC-7 [email protected] wrote: > >> Hey all. >> There must be a simple training GUI for Tesseract? >> Looked but could not find. >> >> Saw mentioned jTessBoxEditor but the download link yields VietOCR, >> a software which seems to do nothing at all. >> >> Any ideas for a simple training GUI? >> >> Software should not be this difficult people. >> It's 2023, not 1982. >> >> I look forward to hearing any ideas you might have. >> > -- You received this message because you are subscribed to the Google Groups "tesseract-ocr" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/73ee10c8-263f-4f46-ba7d-af86906722f3n%40googlegroups.com.
