Choosing a Defect Classifier – Defect Clinic EP.02
Catching NGs but not the cause? Pick IntAnomaly for few samples, IntClassification for known types, IntDetection for location, IntSegmentation for area.
EP.02 · Defect classification · Catching NGs but Not the Cause? How to Choose a Defect Classifier
The inspection system catches NGs just fine, but nobody knows why they happened. To fix the process, defects have to be sorted by cause.
The inspection only separates OK from NG; it does not sort defects by cause. Which engine fits depends on how many defect samples you have and whether you need the type, the location and count, or the area and shape.
Line 3 prescription: IntAnomaly gates OK/NG, then IntClassification sorts only the NGs by cause. Unseen defects go to Unclassified and become the next training data.
Chapters: 0:00 Opening · Line issue / 0:19 How to choose a defect classifier / 0:33 Two-stage pipeline in IntLearn Pro / 0:44 Sorting by cause on the line
Can I apply the fix from a video to my line as is? Each video is a short animation that walks through a problem and a fix. Whether it works on your line depends on a review of samples, imaging conditions and the process.
Every Tuesday 11:00 KST (02:00 UTC), a new episode premieres on YouTube.
* Every episode is an animated dramatization for illustration, and the character voices are AI-synthesized. Numbers, coordinates and charts on screen are examples.
- 1. Few defect samples, unseen defects in the mix: IntAnomaly (anomaly detection), which learns from good parts only.
- 2. Known defect types: IntClassification (classification).
- 3. Need location and count: IntDetection (detection).
- 4. Need area and shape: IntSegmentation (segmentation; CaviarSeg in the video).