Guides to synthetic defect data for machine vision inspection
Guides
- Why use synthetic data for defect detection?
Because the examples you need most are the ones your line makes least.
- Synthetic defect data vs data augmentation
Augmentation varies the defects you already have. Synthetic generation makes ones you don’t have.
- How many defect images do you need to train an inspection model?
What you need depends on your defects, your part and your model, and you can measure it.
- How do you know a synthetic defect label is correct?
Visynex labels every defect image by what a separate detection pass actually finds, and discards the images where it can’t find the defect ordered.
- What is a golden sample, and is one photo enough?
A golden sample is a part known to be within specification, used as the reference that other parts are compared against.
- Synthetic YOLO dataset for defect detection
You receive a defect detection dataset in YOLO and COCO format, ready to read into your own visual inspection training data pipeline.
- Anomaly detection vs supervised defect detection, from one photograph
Two common ways to teach a camera to find defects differ in what they need from you, and in what they tell you back.
- Process-based, procedural or CAD-rendered synthetic defects?
There are three common ways to make synthetic images of defects for training an inspection model.
- Laser-cut sheet metal defect dataset
You send one photograph of a good laser-cut part.
- Synthetic casting porosity images
Porosity is one of the defects a casting line most needs to catch, and one of the hardest to collect examples of.
Measurements: how to check synthetic labels
- How do I check synthetic labels? First, test the checkHow to check synthetic defect labels automatically
Test the check before you trust it: plant a fault it must miss and one it must catch.
- Checking synthetic labels against the reference photo: measure the floor firstHow to check a synthetic defect image by comparing it with the reference photo
A reference comparison carries a floor: 3.83 grey levels at the median, everywhere outside the defect.
- Checking synthetic labels: the same image scored 0.18, then 0.72Why a check’s answer can depend on the other images in the folder
The same check on the same image scored 0.176 one time and 0.719 the next.
Try it on your own part
Sign in with your work email, upload one photograph of a good part, and get up to 10 images of it back, labelled in YOLO and COCO. Each takes about a minute. Free, once per email address.
Try it freeTalk to an engineer
Thirty minutes with someone who built the pipeline. No upload, no account.
Book a call