Splitting PDF File - Confidence Score

Hello,

At my workplace, I need to scan multiple loose sheets using a scanner. Currently, these are saved as a single PDF file. However, these loose sheets belong together. For example, pages 1-5 make up one folder, and pages 6-10 make up another. In theory, I would need to perform two separate scans to ultimately get two PDF files. The first page of each folder is always the same, with the only difference being the folder number. I’ve created a flow that works 90% of the time using AiBuilder and PDF Services from Adobe.

**Example:**
We have a scan process with the following pages: `ABCDEABCDE`. The desired result is two PDF files:
- PDF 1: `ABCDE`
- PDF 2: `ABCDE`.

I created a flow that identifies the position of the first `A`. Then, it checks for the second `A` and uses that to extract the first PDF file. I repeat this iteratively for other documents included in the scan.

Now, here’s what I don’t understand: When I tell the AI Builder to give me the page number of the document it recognizes, what is it referring to? The first page it detects? Or the page where it has the highest confidence score? Currently, it seems to always refer to the first page, which is ideal for me. However, I’ve had the impression in the past that it skips pages if a later page has a higher confidence score. I want to make that decision myself and ensure it always selects the first page. And I want to tell him manually when to skip a single page if the score is too low.

Now let’s consider the following scenario with 3 `A`s: `ABCDEABCDEABCDE`. I split the first `ABCDE`, leaving `ABCDEABCDE`. If I process this new document through the AI Builder again, it assigns the last `A` a confidence score of 20%, and the process takes 20 minutes, so I stop it manually. How can I instruct the AI Builder to proceed even with a confidence score of 20%? That’s sufficient for me. However, I don’t see an option to configure this.

Categories:

AI Builder

Quick Links

Leaderboard