Hello,
I’ve created a document AI processor and I am struggling on how to properly set up the schema for a form that has several checkboxes. I’ve attempted to do this by labeling each checkbox section and create an “optional multiple” field in hopes to extract all selected checkboxes.
For example, there is a section titled Coexisting Conditions, with 34 conditions listed with their own respective checkbox. To extract the conditions that are checked do I need to create 34 fields in the schema with each condition or will something like the below work with enough training?
Here is an example of the form. Any help on how to setup the schema to properly extract the checked options would be greatly appreciated!
Hi @chrishooks22 ,
Welcome to Google Cloud Community!
The approach of creating 34 distinct fields in the schema is likely the most dependable method for extracting checked options from your form due to its clear structure.
Here’s why:
1. Clarity and Structure:
- Direct Mapping: Each schema field corresponds directly to a specific checkbox, ensuring accurate data extraction.
2. Training Efficiency:
- Focused Learning: Separate fields enable your AI to identify specific patterns for each checkbox, improving both training accuracy and efficiency.
3. Flexibility:
- Customizable Extraction: Individual fields provide enhanced adaptability for different layouts or sections of the form.
The extraction process is simplified and reliability is improved by using distinct fields. Therefore, instead of a combined “optional multiple” field, I recommend creating 34 separate fields for the Coexisting Conditions section. This approach will ultimately enhance clarity and accuracy.
I hope the above information is helpful.
2 Likes
