Inspection using ML
Antonio Grimaldo
Why?
- Machine learning algorithms are not common
- Traditional methods require a lot more time to develop a solution and more experience
- Traditional methods don't work well with textures or patterns
The Challenge
Machine vision lab
Chocolates
The challenge
The dataset consisted of 50 images:
- 41 images of "non-defective" chocolate and 9 images of "defective" chocolate
How it works?
Image
Model
Output
SVM
-1 or 1
Chocolate
techniques
- One class SVM
- One class SVM using PCA
- One class SVM with cropped images
- One class SVM with cropped images and PCA
- Isolation trees with cropped images (Bonus)
- One class SVM with cropped and filtered images (Bonus)
Isolation Forest OneClass SVM
Support Vector MAchines
How can we separate this?
High dimensionality!!
Principal Component Analysis
- It is a procedure for identifying a smaller number of uncorrelated variables, called "principal components", from a large set of data.
- It reduces the dimensionality of data
Results
Filtered Image
Average of all images
TP=40 | FP=0 |
---|---|
FN=1 | TN=9 |
One Class SVM with cropped images
Accuracy: 98%
FiN.
ml-inspection
By Antonio Grimaldo
ml-inspection
- 647