AI Inspection With 3 Defect Samples – Defect Clinic EP.06
Only three defect samples? IntAnomaly builds its baseline from good parts and flags what differs. The three samples only check the threshold.
EP.06 · Getting started · Only 3 Defect Samples? AI Inspection That Learns From Good Parts
AI inspection is on hold because there are only three defect samples.
There are too few samples to teach it defects. CEO Na flips it around: teach the AI what normal looks like.
Start with normal.
Chapters: 0:00 Opening · Only three defect samples? / 0:20 Judge lens · IntAnomaly learns from good parts / 0:33 Can it catch new defects? · Validating with three samples / 0:48 Learn normal → classify defects → autonomous control
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. CEO Na's voice is an AI clone of the CEO's own, used with consent; other voices are also AI-synthesized and the music is AI-generated. On-screen values and digital-twin scenes are examples. IntLearn ML autonomous control is under development and needs on-site validation.
- 1. IntAnomaly (the anomaly-detection engine in IntLearn Pro) builds its baseline from good samples only.
- 2. Parts that differ from the baseline get a higher anomaly score, and the odd spot glows red on a heatmap.
- 3. Even defect shapes it has never seen are caught, as long as they differ from normal.
- 4. The three defect samples aren't for training. They're for validation (checking the threshold).
- 5. Once defects pile up, sort them by type with IntClassification, then extend to autonomous control with IntLearn ML (EP.05).