liuruoze

EasyPR 车标识别

4

能够识别车子的类型，例如大众，宝马，奔驰，尼桑，丰田等。

用四周的线信息进行二次精准定位

8

目前定位的车牌，有时候会由于soble或颜色的误差，导致车牌定位不正。或者定位的范围过小，没有覆盖全部的车牌，或者定位的范围过大，只有部分是车牌。如果在一次定位的基础上，再配合车牌四周本身的线信息进行二次定位，就会大幅度减少这种定位不正的现象，从而形成精准定位。 此改进建议放到字符识别完善以后进行。

FCN.py runs right, but FCN_plus.py runs failure

2

This failure is this: `File "FCN_plus.py", line 295, in tf.app.run() File "/home/gpu/liuruoze/Matting/venv2/local/lib/python2.7/site-packages/tensorflow/python/platform/app.py", line 43, in run sys.exit(main(sys.argv[:1] + flags_passthrough)) File "FCN_plus.py", line 181, in main train_dataset_reader = BatchDatset('data/trainlist.mat') File "/home/gpu/liuruoze/Matting/portrait_plus.py",...

line 26: ``` img_data = skio.imread('data/images_data/'+name+'.jpg') ``` May be the x and y mask should come from "images_data_crop" folder, according to the paper.

How to detect that the opponent computer says 'gg' and accept that?

2

When training the agent, we find that sometimes the opponent says 'gg' which means we win. Is there a way for the agent to accept this surrender now? Thank you...

This pull request fix one bug and close 2 issues.

1

Hi, this pr will fix one bug in discrete.py, make the example work correctly, and close 2 issues which are #4 and #5 . I have test this code in...