Person-Segmentation-Keras

Person-Segmentation-Keras copied to clipboard

reshape to get sharp masks

3

I noticed that when I change the reshape from https://github.com/TianzhongSong/Person-Segmentation-Keras/blob/master/models/unet.py#L71 into: conv11 = Reshape((input_height * input_width, nClasses))(conv11) the UNet seems to remove the line artifacts seen in the output masks.

ldenoue

eval_segmentation.py

6

@TianzhongSong，您好！ gt_img[:] += (gt[:] == 1).astype('uint8') pt_img[:] += (pt[:] == c).astype('uint8') gt_img[:] shape=（1，50176， 2） gt[:] shape=（50176，） pt_img[:] shape=（1，50176， 2） pt[:] shape=（50176，） 这应该也是个问题吧

buaa-luzhi

Traceback (most recent call last): File "predict.py", line 47, in predict_segmentation() File "predict.py", line 20, in predict_segmentation raise ValueError('Do not support {}'.format(args.model)) ValueError: Do not support 'unet Please tell me...

123lc-123

I can't download the dataset

3

Hello how can i download the dataset if i'm not from china and i don't have a baidu account. any help please?

M-Ghorbel

Was it because I was training less ??

WuGe0424

Can we download a pre-trained model?

JaeDukSeo

error : python train_segmentation.py --model='unet'

1

dear, Epoch 1/50 Traceback (most recent call last): File "train_segmentation.py", line 102, in main() File "train_segmentation.py", line 90, in main callbacks=[checkpoint], File "/home/vsmart/anaconda3/envs/tensorflow/lib/python3.7/site-packages/keras/legacy/interfaces.py", line 91, in wrapper return func(*args, **kwargs)...

NguyenDangBinh

dear, Can you show me the tree of data? Thank you!

NguyenDangBinh

Hello, I ran 50 epochs according to the code, but it did not achieve satisfactory results. How many epochs do I need to train to achieve such an effect?

Hazeliii

Pythonlushun