Update README.md

[caoyunkang]

Hi! Thanks for your inspiring survey on industrial anomaly detection. I have recently published a paper addressing point cloud anomaly detection. If possible, I would like to add this paper to your repo and survey. Additionally, I have been devoted myself to industrial anomaly detection, I'll appreciate it a lot if you could discuss with me whenever encountered problems related to this topic.

Thanks!

[guoyang-xie]

Thanks a lot! We have merged your request

[danielsoy]

Hi, I am interested in IAD too!

Look this video please:

https://www.youtube.com/watch?v=W4-ZtpJtE5c&ab_channel=tektronix475

As you can see, the anomaly score for the pasta sample is higher when the black spot(defect) is located on the right side of the frame. With a mean score of around 25 , and a mean score of 20 if the anomaly, is on the left side.

The model is Padim from resnet_18, and was take from: https://github.com/OpenAOI/anodet.

Here you can download the dataset and the inferred images results n the training dataset.(250Mb).

https://drive.google.com/drive/folders/1_OF6-_MNTskgPd5Sqqtm1-p4WqDu0Jhl?usp=share_link

In the folders with the results gotten with Patchcore, there is almost no score difference between the inferred samples with the black spot on the left and the ones with the anomaly on the right side.

Whats your take about this issue?

Thanks a lot!

[guoyang-xie]

Dear Danielsoy:

Well done!

If you desire to open public your dataset in our repo, could you give more information for your dataset?

Be more specific for the description of this dataset.

Thanks a lot!

Guoyang Xie PhD Candidate of University of Surrey

On Apr 21, 2023, at 02:20, danielsoy @.***> wrote:

Hi, I interested in IAD too!

Look this video please:

https://www.youtube.com/watch?v=W4-ZtpJtE5c&ab_channel=tektronix475

As you can see, the anomaly score for the pasta sample is higher when the black spot(defect) is located on the right side of the frame. With a mean score of around 25 , and a mean score of 20 if the anomaly, is on the left side.

The model is Padim from resnet_18, and was take from: https://github.com/OpenAOI/anodet.

Here you can download the dataset and the inferred images results n the training dataset.(250Mb).

https://drive.google.com/drive/folders/1_OF6-_MNTskgPd5Sqqtm1-p4WqDu0Jhl?usp=share_link

— Reply to this email directly, view it on GitHub https://github.com/M-3LAB/awesome-industrial-anomaly-detection/pull/1#issuecomment-1516758696, or unsubscribe https://github.com/notifications/unsubscribe-auth/AWU3OGZMRWFRNVVIOP2ESNTXCF47XANCNFSM6AAAAAAWIW5UUE. You are receiving this because you commented.