yeahhuang / al_surface_defect_detection Goto Github PK

图片直接定位 预测？

配置文件是哪个？

我终于把那些钢管的图片都转换成了COCO的json格式，已经在跑了。不过1080ti占了8个G显存。谢谢你告诉我可以在极客云上租服务器！

Hi, I has installed Detectron on my Ubuntu 16.04 with cuda and cudnn, but I don't know how to run your codes?
Thank you!

你好，请问您是否还有该代码的数据集呢？比赛已经过了报名时间不能再下载数据了。

有没有数据分享一下呢？

Thnx，your best work！！！i meet a question that dataset need to be calibrated by myself in Season2(localization) ？？？？
thnx so much to reply this question！

同学，这个数据集你还留着么，跪求这个的数据集呀

Is there already dataset inside the folder? If not, where can I put the data?
Which file should I run first? Can you also show the steps to train and test the data? Thanks!

你好，对你的工作表示非常感兴趣，请问可以分享一下数据集吗？谢谢，我的邮箱[email protected]

链接：https://pan.baidu.com/s/1R2RW-ql0UoO_m5C0X_BYNA 提取码：w7tn

由于我这边只需要分类好坏而不需要将缺陷具体标出，感觉和初赛的题目比较类似。问一下你初赛提交的模型是什么呢？似乎项目里没有上传，请问是有官方版本之类的可以下载吗？谢谢！

how to deal with dataset?
how to run this season2 project?(steps)
plz!!!,thx ur sharing!

As your blog's description, OHEM is more important part to improve mAP, but I could not find any implement from all winner‘s release? Do you known any runnable project?

Hello, @YeahHuang
I train my own dataset with the config file"tutorial_1gpu_e2e_faster_rcnn_R-50-FPN.yaml", and the result of test_net.py is :

INFO voc_dataset_evaluator.py: 145: Mean AP = 0.6124
INFO voc_dataset_evaluator.py: 146: TP = 6802
INFO voc_dataset_evaluator.py: 147: FP = 3439
INFO voc_dataset_evaluator.py: 148: FN = 1388
INFO voc_dataset_evaluator.py: 149: Mean 1-Precision = 0.3358
INFO voc_dataset_evaluator.py: 150: Mean Recall = 0.8305

I want to increase the recall as far as possible.
And my dataset is similar to AI_Surface_defect. The classes is 5 + 1(background).

Could you give me some advices? thanks so much!
tutorial_1gpu_e2e_faster_rcnn_R-50-FPN.txt

How to handle the “无瑕疵图片”？
Thank you！

我的AP实在是太低太低太低太低了，哭了

 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.048
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.109
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.030
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.108
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.070
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.058
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.228
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.307
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.307
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.117
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.147
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.327
INFO json_dataset_evaluator.py: 218: Wrote json eval results to: ./test_result/test/defect_val/generalized_rcnn/detection_results.pkl
INFO task_evaluation.py:  62: Evaluating bounding boxes is done!
INFO task_evaluation.py: 181: copypaste: Dataset: defect_val
INFO task_evaluation.py: 183: copypaste: Task: box
INFO task_evaluation.py: 186: copypaste: AP,AP50,AP75,APs,APm,APl
INFO task_evaluation.py: 187: copypaste: 0.0478,0.1087,0.0303,0.1079,0.0703,0.0578