search

ShuaiLYU / Deep-Learning-Approach-for-Surface-Defect-Detection

A Tensorflow implementation of "Segmentation-Based Deep-Learning Approach for Surface-Defect Detection"
730 stars 258 forks source link

How to prepare data #1

Open PowerDi opened 5 years ago

PowerDi commented 5 years ago

Your work is outstanding. But could u tell me how should I prepare the image? Does it need label the defects firstly with bounding box or pixel segmentation?

ShuaiLYU commented 5 years ago

You should label the defects with pixel segmentation.

Shahdoust commented 5 years ago

@Wslsdx What tool did you use for pixel segmentation or what tool is good to use here?

ShuaiLYU commented 5 years ago

@Shahdoust what dose "pixel segmentation" mean?

Shahdoust commented 5 years ago

@Wslsdx thank you for your reply. Sorry I am a little confused and new here. I meant the exact tool like LabelMe or PixelAnnotationTool that provides to segment your images manually. I really appreciate if you help me to prepare my data.

ShuaiLYU commented 5 years ago

@Shahdoust LabelMe or PixelAnnotationTool is oK. I prepare my data with Halcon .

nys09 commented 4 years ago

@Wslsdx Does Halcon have pixel-level labeling capabilities? If not, what is the best way to label images at pixel level?

li675599061 commented 4 years ago

@Shahdoust So what tools did you end up using for data tagging?Which tool is effective?Can you tell me?Thank you very much

Shahdoust commented 4 years ago

@li675599061 I think the PixelAnnotation tool is good and easy to use.

qwfeng commented 3 years ago

@Shahdoust ,windows10安装和使用 PixelAnnotation tool的方法教程有嘛

qwfeng commented 3 years ago

windows使用Halcon制作数据集的教程有嘛