unet
unet copied to clipboard
zhixuhao
用自己的数据训练后，输出的预测结果全黑
不知道咋搞的，预测结果输出来是全黑的，我乘255以后就是灰色的了，看了下值，都是在127左右，怎么也想不通究竟哪里有问题，大家有遇到这个问题的吗？能给个方向吗？我感觉是数据有问题，但不知道怎么入手
输出的是 sigmoid 结果，值域范围[0,1]，做一下处理就好
y_pred = model.predict(image)
y_pred = (y_pred > 0.5).astype(np.uint8).reshape(IMAGE_H, IMAGE_W)
plt.imshow(np.squeeze(y_pred), plt.cm.gray)
请参考我的代码 TFLearn U-Net Starter https://github.com/SixQuant/U-Net/blob/master/kernel.ipynb
不知道咋搞的，预测结果输出来是全黑的，我乘255以后就是灰色的了，看了下值，都是在127左右，怎么也想不通究竟哪里有问题，大家有遇到这个问题的吗？能给个方向吗？我感觉是数据有问题，但不知道怎么入手
问一下，自己的数据需要怎样的处理。我用自己数据跑会出现维度问题ValueError: Error when checking : expected input_1 to have 4 dimensions, but got array with shape (1, 256, 256, 3, 1)
不知道咋搞的，预测结果输出来是全黑的，我乘255以后就是灰色的了，看了下值，都是在127左右，怎么也想不通究竟哪里有问题，大家有遇到这个问题的吗？能给个方向吗？我感觉是数据有问题，但不知道怎么入手
问一下，自己的数据需要怎样的处理。我用自己数据跑会出现维度问题ValueError: Error when checking : expected input_1 to have 4 dimensions, but got array with shape (1, 256, 256, 3, 1)
你多个了一个维度，读出来的图片应该是(256, 256, 3)的，不知道你咋就搞成(1, 256, 256, 3, 1)了。不过没关系，reshape 一下 reshape(1, 256, 256, 3)
不知道咋搞的，预测结果输出来是全黑的，我乘255以后就是灰色的了，看了下值，都是在127左右，怎么也想不通究竟哪里有问题，大家有遇到这个问题的吗？能给个方向吗？我感觉是数据有问题，但不知道怎么入手
问一下，自己的数据需要怎样的处理。我用自己数据跑会出现维度问题ValueError: Error when checking : expected input_1 to have 4 dimensions, but got array with shape (1, 256, 256, 3, 1)
你多个了一个维度，读出来的图片应该是(256, 256, 3)的，不知道你咋就搞成(1, 256, 256, 3, 1)了。不过没关系，reshape 一下 reshape(1, 256, 256, 3)
是啊，我用这个代码跑rgb图的，多了后面那个维度，然后改了一下代码，定义geneTrainNpy函数时少加一维就可以跑了
个人建议：（1）考虑设置不同的随机初始化种子，比如seed=0或者1；（2）考虑在卷积操作之后，增加batch normalization 或者 dropout.
You can try more epochs to improve the acc . Another place to check is the function of adjustData.
def adjustData(img,mask,flag_multi_class,num_class):
#here#
elif(np.max(img) > 1):
img = img / 255
mask = mask /255
mask[mask > 0.5] = 1
mask[mask <= 0.5] = 0
return (img,mask)
This code works for py3, if you use py2, modify '255' to '255.'; Last, you need to check the threshold '0.5' in order to make sure the returned mask is a binary(0/1) mask.
楼上的各位大佬，有个问题想咨询一下。
我在博主的代码里看到好多DROPOUT层，但是从网络的结构图上没看到，请问这些是自己加得吗？
还有个问题，从网络结构图上看最后一个箭头代表1*1的卷积核，但是为什么博主的代码里是
conv9 = Conv2D(2, 3, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(conv9) conv10 = Conv2D(1, 1, activation = 'sigmoid')(conv9)
而不是 conv9 = Conv2D(2, 1, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(conv9) ，十分感谢！
楼上的各位大佬，有个问题想咨询一下。 我在博主的代码里看到好多DROPOUT层，但是从网络的结构图上没看到，请问这些是自己加得吗？ 还有个问题，从网络结构图上看最后一个箭头代表1*1的卷积核，但是为什么博主的代码里是
conv9 = Conv2D(2, 3, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(conv9) conv10 = Conv2D(1, 1, activation = 'sigmoid')(conv9)而不是conv9 = Conv2D(2, 1, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(conv9)，十分感谢！
dropout层是为了防止过拟合才加上去的，如果你训练出来的模型没有过拟合的情况，可以不用加上这些层
不知道咋搞的，预测结果输出来是全黑的，我乘255以后就是灰色的了，看了下值，都是在127左右，怎么也想不通究竟哪里有问题，大家有遇到这个问题的吗？能给个方向吗？我感觉是数据有问题，但不知道怎么入手
我也遇到了相同的问题，确实是数据的问题，首先训练的数据必须是tif格式图片，不然出来的结果肯定是全黑的，还有图片的大小也有影响。 我自己改过之后的代码你可以参考一下：https://github.com/wuyang0329/unet
ｈｉ @sanersbug ,
Do you resolve this problem now? if you are done, please share some details. Thanks.
ｈｉ @sanersbug ,
Do you resolve this problem now? if you are done, please share some details. Thanks.
I have resloved this problem ，you can see this：https://github.com/wuyang0329/unet
hi @wuyang0329 , When I tried to run your code and original data, I faced below problem...do you know how to fix it?
Traceback (most recent call last):
File "/home/dongs/Documents/Software/pycharm-community-2018.2.4/helpers/pydev/pydevd.py", line 1664, in
hi @wuyang0329 , When I tried to run your code and original data, I faced below problem...do you know how to fix it?
Traceback (most recent call last): File "/home/dongs/Documents/Software/pycharm-community-2018.2.4/helpers/pydev/pydevd.py", line 1664, in main() File "/home/dongs/Documents/Software/pycharm-community-2018.2.4/helpers/pydev/pydevd.py", line 1658, in main globals = debugger.run(setup['file'], None, None, is_module) File "/home/dongs/Documents/Software/pycharm-community-2018.2.4/helpers/pydev/pydevd.py", line 1068, in run pydev_imports.execfile(file, globals, locals) # execute the script File "/home/dongs/Documents/Software/pycharm-community-2018.2.4/helpers/pydev/_pydev_imps/_pydev_execfile.py", line 18, in execfile exec(compile(contents+"\n", file, 'exec'), glob, loc) File "/home/dongs/Documents/Project/unet-road/main.py", line 24, in results = model.predict_generator(test_data,steps=20,verbose=1) File "/home/dongs/anaconda3/lib/python3.6/site-packages/keras/legacy/interfaces.py", line 91, in wrapper return func(*args, **kwargs) File "/home/dongs/anaconda3/lib/python3.6/site-packages/keras/engine/training.py", line 1522, in predict_generator verbose=verbose) File "/home/dongs/anaconda3/lib/python3.6/site-packages/keras/engine/training_generator.py", line 435, in predict_generator generator_output = next(output_generator) File "/home/dongs/anaconda3/lib/python3.6/site-packages/keras/utils/data_utils.py", line 709, in get six.reraise(*sys.exc_info()) File "/home/dongs/anaconda3/lib/python3.6/site-packages/six.py", line 693, in reraise raise value File "/home/dongs/anaconda3/lib/python3.6/site-packages/keras/utils/data_utils.py", line 685, in get inputs = self.queue.get(block=True).get() File "/home/dongs/anaconda3/lib/python3.6/multiprocessing/pool.py", line 670, in get raise self._value File "/home/dongs/anaconda3/lib/python3.6/multiprocessing/pool.py", line 119, in worker result = (True, func(*args, **kwds)) File "/home/dongs/anaconda3/lib/python3.6/site-packages/keras/utils/data_utils.py", line 626, in next_sample return six.next(_SHARED_SEQUENCES[uid]) File "/home/dongs/Documents/Project/unet-road/data.py", line 110, in testGenerator img = io.imread(os.path.join(self.test_path,filename),as_gray=True) File "/home/dongs/anaconda3/lib/python3.6/site-packages/skimage/io/_io.py", line 61, in imread img = call_plugin('imread', fname, plugin=plugin, **plugin_args) File "/home/dongs/anaconda3/lib/python3.6/site-packages/skimage/io/manage_plugins.py", line 211, in call_plugin return func(*args, **kwargs) File "/home/dongs/anaconda3/lib/python3.6/site-packages/skimage/io/_plugins/tifffile_plugin.py", line 31, in imread return tif.asarray(**kwargs) TypeError: asarray() got an unexpected keyword argument 'as_gray'
maybe you should check your skicit-image’s version. Mine is 0.14.1
hi @wuyang0329 , When I tried to run your code and original data, I faced below problem...do you know how to fix it?
Traceback (most recent call last): File "/home/dongs/Documents/Software/pycharm-community-2018.2.4/helpers/pydev/pydevd.py", line 1664, in main() File "/home/dongs/Documents/Software/pycharm-community-2018.2.4/helpers/pydev/pydevd.py", line 1658, in main globals = debugger.run(setup['file'], None, None, is_module) File "/home/dongs/Documents/Software/pycharm-community-2018.2.4/helpers/pydev/pydevd.py", line 1068, in run pydev_imports.execfile(file, globals, locals) # execute the script File "/home/dongs/Documents/Software/pycharm-community-2018.2.4/helpers/pydev/_pydev_imps/_pydev_execfile.py", line 18, in execfile exec(compile(contents+"\n", file, 'exec'), glob, loc) File "/home/dongs/Documents/Project/unet-road/main.py", line 24, in results = model.predict_generator(test_data,steps=20,verbose=1) File "/home/dongs/anaconda3/lib/python3.6/site-packages/keras/legacy/interfaces.py", line 91, in wrapper return func(*args, **kwargs) File "/home/dongs/anaconda3/lib/python3.6/site-packages/keras/engine/training.py", line 1522, in predict_generator verbose=verbose) File "/home/dongs/anaconda3/lib/python3.6/site-packages/keras/engine/training_generator.py", line 435, in predict_generator generator_output = next(output_generator) File "/home/dongs/anaconda3/lib/python3.6/site-packages/keras/utils/data_utils.py", line 709, in get six.reraise(*sys.exc_info()) File "/home/dongs/anaconda3/lib/python3.6/site-packages/six.py", line 693, in reraise raise value File "/home/dongs/anaconda3/lib/python3.6/site-packages/keras/utils/data_utils.py", line 685, in get inputs = self.queue.get(block=True).get() File "/home/dongs/anaconda3/lib/python3.6/multiprocessing/pool.py", line 670, in get raise self._value File "/home/dongs/anaconda3/lib/python3.6/multiprocessing/pool.py", line 119, in worker result = (True, func(*args, **kwds)) File "/home/dongs/anaconda3/lib/python3.6/site-packages/keras/utils/data_utils.py", line 626, in next_sample return six.next(_SHARED_SEQUENCES[uid]) File "/home/dongs/Documents/Project/unet-road/data.py", line 110, in testGenerator img = io.imread(os.path.join(self.test_path,filename),as_gray=True) File "/home/dongs/anaconda3/lib/python3.6/site-packages/skimage/io/_io.py", line 61, in imread img = call_plugin('imread', fname, plugin=plugin, **plugin_args) File "/home/dongs/anaconda3/lib/python3.6/site-packages/skimage/io/manage_plugins.py", line 211, in call_plugin return func(*args, **kwargs) File "/home/dongs/anaconda3/lib/python3.6/site-packages/skimage/io/_plugins/tifffile_plugin.py", line 31, in imread return tif.asarray(**kwargs) TypeError: asarray() got an unexpected keyword argument 'as_gray'
Maybe you are using skimage version 0.13, you can see help(skimage.io.imread) and find the argument is actually as_grey, hope this might help
我自己的训练数据是rgb的，这个项目默认的数据是灰度的，所以有些地方要改，参照这个问答就可以了，https://github.com/zhixuhao/unet/issues/59
我也遇到类似的情况，我把网络改成1维的了，训练集每一个样本里大概2000个0，10个1，我训练出来结果大概都是0.47，我考虑可能是正负样本分布不均造成的，我给交叉熵加入了权重，然后出现了过拟合现象，在测试集上都是错的，后来我有把代码该成每一个epoch洗牌一次，可是这样真的就学习不到特征了，很苦恼，我也不知道怎么解决。
