Just getting started, I would like to consult this algorithm.

[cylvzj]

@skokec

Hello, can you tell me about it?

1. Accuracy of prediction

2. Time-consuming prediction of each picture

3. And your GPU configuration.

4. The difference from this article.

Https://github.com/Wslsdx/Deep-Learning-Approach-for-Surface-Defect-Detection

No embarrassment, just study, thank you!

[skokec]

Hi, the linked github code is an unofficial 3th party code for the same model as we presented in our paper. I do not know the difference between theirs and our implementation.

For more information about the accuracy of our model see our paper: Segmentation-Based Deep-Learning Approach for Surface-Defect Detection.

On KolektorSDD images (around 1500x500 px) this algorithm takes around 120-150 ms for the inference. See our ICVS2019 conference paper Deep-learning-based computer vision system for surface-defect detection for more details.

[cylvzj]

Thank you for your reply. Please forgive me for my poor English.

------------------ 原始邮件 ------------------ 发件人: "Domen Tabernik"[email protected]; 发送时间: 2019年7月25日(星期四) 晚上8:44 收件人: "skokec/segdec-net-jim2019"[email protected]; 抄送: "A 世智智能科技"[email protected];"Author"[email protected]; 主题: Re: [skokec/segdec-net-jim2019] Just getting started, I would like toconsult this algorithm. (#4)

Hi, the linked github code is an unofficial 3th party code for the same model as we presented in our paper. I do not know the difference between theirs and our implementation.

For more information about the accuracy of our model see our paper: Segmentation-Based Deep-Learning Approach for Surface-Defect Detection.

On KolektorSDD images (around 1500x500 px) this algorithm takes around 120-150 ms for the inference. See our ICVS2019 conference paper Deep-learning-based computer vision system for surface-defect detection for more details.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.