image-segmentation-keras

image-segmentation-keras copied to clipboard

I was looking at "image_segmentation_generator" in the data_loader.py and printed out the shape of the individual output image and mask shapes. I got: img shape (416, 608, 3) seg_shape (252928,...

mavaylon1

Changing metrics in model.train()

1

Hello there, first of all thank you for this amazing package. I was wondering if it is possible to change the metric used during the model.train stage. In the source...

santiactis

would you mind providing more information about the provided dataset1

1

hello, would you mind providing more information about the segmentation dataset , such as the practical map from label index to label? Thanks

miss1997yuan

Prediciton: Inference resolution vs Output resolution

3

Calling a prediction of a model gives wo different resolutions. ``` model = vgg_unet(n_classes=2 , input_height=1280, input_width=640 ) model.load_weights('myweights1280x640') outcome=model.predict_segmentation( inp='input.png', out_fname="out.png") ``` this results in the array "outcome" wich...

GGDRriedel

Where are segnet maxpool indices leveraged ?

1

Hi all, Very interesting blog article and codebase, thank you. I was interested in trying to compare the various segmentation strategies mentionned in the blog article, and realized that the...

gnthibault

no way to train pascal voc 2012 dataset

1

@divamgupta I have tried many models to train from scratch with pascal voc 2012 datasets. However, I could not get a good trained model to segment the image. Is there...

white2018

Forecast result graph

1

Excuse me, I have a question to ask you.Why are there so many gradients in the predicted result？And it’s not the same color as the label image.

TGUCJX

Upsampling in Decoder of SegNet

1

Hi all, I was checking decoder implementation of SegNet and saw that Upsampling2D is used [here](https://github.com/divamgupta/image-segmentation-keras/blob/bb7ac1f5cbaa884cea94e56c1a2aa5ea1690ed26/keras_segmentation/models/segnet.py#L21). Keras provides bilinear and nearest upsampling for Upsampling2D according to the [docs](https://www.tensorflow.org/api_docs/python/tf/keras/layers/UpSampling2D), so one...

ilteralp

It seems to me that before many convolution layers, the Zeropadding 2d is applied to keep the original dimension of the input. I think one leaner way to implement is...

bobvo23

Will this be update to use .fit instead of fit_generator? I tried and only works when I keep it at fit_generator.

mavaylon1