Hi everyone,
I have a question about OCR.
In the image I want to extract data from, there are two sections. One contains static data (OCR reads it without problems), and the second part is where I’m having trouble. We could call it a subtable; it’s like saying the first section contains information about an inspection and the second contains information about the findings. The problem is that, as far as I can see, it can’t be done. Has anyone else had a similar problem?
I believe OCR is not a good option for your use csse.
A) OCR needs training the model, which in a way means , the data pattern to be read has to be consistent in layout. For example a visiting card, an electricity meter etc. So the second part of your data does not seem to be a good use case.
B) OCR is in beta and I believe it has not been extensively used so far.
Alternatively, please take a look at Gemini AI integration in AppSheet.
I’ve been researching Gemini AI and it says I can use JSON along with OCR to store information and have a programmed bot add it to its table. Is this possible, or have you seen anything similar?
To make use of native Gemini feature in AppSheet , please take a look at the following article. I am also sharing the image of the relevant excerpt from the article, which matches your use case.
Excerpt from the above article is in image below for your immediate reference.
However using native Gemini integration feature in AppSheet requires an Enterprise pricing plan.
I had also published a relevant tip of using Gemini with AppSheet to extract information from images. This tip utilizes Google Apps Script(GAS) to integrate Gemini AI with AppSheet. I believe you can use it in core plans as well. Obviously that was a tip on a test app. So you may need more finetuning in all respects to make it viable for you. For example, I used just small used cases such as electric meter to read from images. Here is the tip:
