image-segmentation-keras
image-segmentation-keras copied to clipboard
Implementation of Segnet, FCN, UNet , PSPNet and other models in Keras.
I use the minidataset in `image-segmentation-keras/test/example_dataset` , my acc is: ``` 2020-03-03 18:11:22.262469: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcudnn.so.7 512/512 [==============================] - 109s 212ms/step - loss: 0.3495 - acc:...
Hi, Is there support for VOC data set or COCO dataset when annotating images using tools like labelme Kind Regards.
Can you please provide me the weights file of trained Unet model?
Hi @divamgupta ; Thank you for your great contribution! Everything works fine for me. This repo helps me a lot. However, I noticed that there are only pretrained models for...
Hi, `imageio` dependency is indicated as `imageio==2.5.0` in `setup.py` (which results in down-grading of the latest version of imageio if it is already installed), but in the `requirements.txt` it is...
When I run this code ``` from keras_segmentation.pretrained import pspnet_50_ADE_20K , pspnet_101_cityscapes, pspnet_101_voc12 model = pspnet_101_voc12() model.train( train_images = "train_images/", train_annotations = "train_anno/", checkpoints_path="tmp/pspnet_1", epochs=5 ) ``` I get this...
Why is it the dimension of the mask as follows in data_utils.py seg_labels = np.reshape(seg_labels, (width*height, nClasses)) Why isnt it (batch size, widht, height, nclasses)
Hi, thanks for making this good code available. I have a question about evaluation metrics, in general when we put accuracy without any details like in some paper, we take...
Hi, I am trying to train the model on the dataset for ten classes in the given link http://sceneparsing.csail.mit.edu/, like my query is if I want to train it for...
The model is training and working in general but while predicting I get a bunch of unresolved object warnings, see below. Tensorflow 2.3.1 Model: mobilenet-segnet ``` WARNING:tensorflow:Unresolved object in checkpoint:...