> For the complete documentation index, see [llms.txt](https://captic-2.gitbook.io/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://captic-2.gitbook.io/docs/stich-and-label-check-for-bags-coming-soon.md).

# Seam Inspection

AI vision that checks every bag for badly sewn seams and missing labels, so defective bags don't reach your customers.

Captic Seam Inspection checks every bag on your line for badly sewn seams, and optionally for missing labels, before it leaves your site. For each bag it returns a **pass/fail** result with a **confidence score**, and your line rejects the bad ones.

## Why seams are hard to inspect

Manual inspection is slow, inconsistent and expensive, and a single defective bag that reaches a customer can lead to a claim. Traditional vision systems need to be programmed for every defect type, and struggle with the natural variation in bags, print and sewing.

## What it does

* **Detects badly sewn seams** on every bag, at line speed.
* **Detects missing labels**, as an option.
* **Gives a clear result:** pass or fail, with a confidence score your PLC can act on.
* **Logs every inspection.** Each image and result is stored, giving you full traceability. Review them in [Captic Cloud](/docs/overview/hands-on/user-interfaces.md#captic-cloud).
* **Improves over time.** When something changes on your line, such as a new bag type, new print or different lighting, we retrain the model with your images.

## At a glance

|                 |                                                            |
| --------------- | ---------------------------------------------------------- |
| Inspection time | Under 300 ms per bag                                       |
| Checks          | Seams; labels optional                                     |
| Output          | Pass/fail and confidence score, to your PLC                |
| Camera          | Typically an Advantech ICAM-540, mounted over the conveyor |
| Rejection       | By your line, usually the existing check-weigher           |
| Traceability    | Every image and result logged                              |

## How it works

1. A trigger from your line tells the system a bag is in view.
2. The camera captures the bag.
3. The AI checks the seam, and the label if enabled.
4. The result is sent to your PLC through the [Integration Flow](/docs/overview/hands-on/integrating-your-robot/integration-flow.md).
5. Your line, usually the check-weigher, rejects bags that fail.

## Setting up

Follow [Getting Started](/docs/stich-and-label-check-for-bags-coming-soon/getting-started.md) to go from delivery to the first inspected bag.

## Key settings

* **Region of interest (ROI):** the part of the image that is inspected.
* **Sensitivity:** how strict the inspection is, which trades off missed defects against false rejects.

{% hint style="info" %}
Setting up Seam Inspection on a new line? Contact <support@captic.com> and we'll help you tune it.
{% endhint %}