Using Unet to segmentate LV-MHP-v1 dataset need to trans the LV-MHP-V1 data to .npy file
1 download dataset(LV-MHP-v1 or others)
2 x\y_trans_png.py
base on train.txt /test.txt (filename of train/test data) devide the train/test image and resize
set the y_train/y_test data to black and white data
3 prepare_npy.py
you might need to understand my code and change some.
inputshape(none，128，128，3) outputshape(None, 128, 128, 1)
WARNING!!!!
the label(y_train/y_test) is 0/1 not 0-255 !!!!!!!
QQ 1923269508 wechat: Stephon_Marbury_