Skip to content

This repository aims to measure the zero-shot segmentation performance of Segment Anything Models (SAM) on Industrial defect data

Notifications You must be signed in to change notification settings

Kiminjo/MVTec-SAM-Validation

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

MVTec-SAM-Validation

This repository aims to measure the zero-shot segmentation performance of Segment Anything Models (SAM) on Industrial defect data

Experiment configs

  • prompt type
    • bounding box
    • point
  • data
    • MVTec bottle
    • MVTec screw
    • MVTec wood data

RUN

  1. Get data using dvc
cd data 
dvc pull mvtec.dvc 

  1. Set MVTec's category in main.py
# main.py line 11
category = "wood" # Change here 

  1. Run Code
python main.py

Then the results folder will be created and contain the inference result images.

  1. (Option) Load previously learned results
dvc pull results.dvc 

About

This repository aims to measure the zero-shot segmentation performance of Segment Anything Models (SAM) on Industrial defect data

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages