search

GeorgeSeif / Semantic-Segmentation-Suite

Semantic Segmentation Suite in TensorFlow. Implement, train, and test new Semantic Segmentation models easily!
2.5k stars 880 forks source link

Using matrix instead of image #217

Open pavva94 opened 5 years ago

pavva94 commented 5 years ago

Hi, Thanks sooo much for this repo it's very very helpful.

My question is about use as dataset matrix of number (range: 0-60000) instead of image. I know that you can tell me to convert the matrix into image but the result are very bad and at the end I need to have a matrix of number. I know that images are matrix of pixel rgb but I haven't a conversion for my data into rgb.

So, how can I give in input a matrix? if it's possibile.

My idea of input image are matrix 0-60000 and label image is 0-1 for delimit the area to recognize.

Maybe someone helps me.

Thanks