getting Good results for Bottle data, but worse for cable data

plutoyuxie / AutoEncoder-SSIM-for-unsupervised-anomaly-detection-

Improving Unsupervised Defect Segmentation by Applying Structural Similarity to Autoencoders

123 stars 26 forks source link

getting Good results for Bottle data, but worse for cable data #12

Open midhunandu opened 3 years ago

midhunandu commented 3 years ago

bottle jpg above figure shows groundtruths and outputs obtained for bottle data. It seems good.

Untitled above figure shows groundtruths and outputs obtained for cable data. Results are not good. I have used the same arguments as in the readme document for running train and test. Is there any solutions to make it more accurate?

plutoyuxie commented 3 years ago

Hi, @midhunandu Thanks for carefully checking results of our implementation.

As I know, pixel level anomaly detection methods based on RECONSTRUCTION is far behind some new methods nowadays.

We propose a new algorithm. The performance is much better.

nnajeh commented 3 years ago

@plutoyuxie can you present some of the new anomaly detection methods please?

plutoyuxie commented 3 years ago

PaDiM, PatchCore, cflow ad

---Original--- From: @.> Date: Mon, Aug 16, 2021 17:38 PM To: @.>; Cc: @.**@.>; Subject: Re: [plutoyuxie/AutoEncoder-SSIM-for-unsupervised-anomaly-detection-] getting Good results for Bottle data, but worse for cable data (#12)

@plutoyuxie can you present some of the new anomaly detection methods please?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android.