search

tusen-ai / simpledet

A Simple and Versatile Framework for Object Detection and Instance Recognition
Apache License 2.0
3.08k stars 486 forks source link

How to draw detection frame, category, and confidence on test set ? #314

Closed dongzhenguo2016 closed 4 years ago

dongzhenguo2016 commented 4 years ago

I now run detection_train.py to train the model using tridentnet_r101v1c4_c5_1x.py, and use detection_test.py to call the trained weight file to calculate the AP values for each class on the test set. Now I want to draw a detection frame and the corresponding category and confidence on each image on the test set. But I didn't find the corresponding code. Can you provide it? Thank you My test set has about 1,000 images. I want to draw the detection frame, category, and confidence. Let ’s take a look at the real detection effect. I see that it is true, only on the test set. In order to see how good the performance of the algorithm is. Thanks a lot!

dongzhenguo2016 commented 4 years ago

please help me! thanke you very much

xchani commented 4 years ago

Sorry, we don't have script to visualize result till now. Since I'm busy recently, this feature may be added not earlier than next week.

dongzhenguo2016 commented 4 years ago

画框已经解决了，我把代码写在了CSDN博客里面，有需要的朋友可以 看一下。 我里面有详细的注释 所能实现的功能就是调用训练好的权重文件，在测试集上画框并显示类别和置信度，还可以指定使用第几个epoch的权重，并且还能将ground truth画上去，方便查看模型真实的效果咋样。 https://blog.csdn.net/dongzichen2015/article/details/104473180 希望大家科研顺利！