search

Justin-Le / hyperspectral_classifier_DBN

A Deep Belief Network for the classification of materials in hyperspectral images.
18 stars 8 forks source link

Need Help #1

Open lotfiGipsyKing opened 7 years ago

lotfiGipsyKing commented 7 years ago

Hello Dear, So i am a phd student in MASCIR Foundation (Moroccan Foundation for Advanced Science, Innovation & Reserarch) , i work in project of hyperspectrale image classification with Deep Learnig , and search for a large Dataset of hyperspectral image and grounf thruth image , because one image of ground-thruth not enought to make a training in a Deep network , i look for your help to find one

Justin-Le commented 7 years ago

Hi,

Multiple large datasets can be found at these two links:

http://www.ehu.eus/ccwintco/index.php?title=Hyperspectral_Remote_Sensing_Scenes https://scien.stanford.edu/index.php/hyperspectral-image-data/

I'm not aware of any others.

Best of luck.

lotfiGipsyKing commented 7 years ago

Thanks , but this Dataset giva only one hyperspectral image an her groundthruth, but in the Deep Learnig training you need a large dataset , you are worked with DBN , isn't ??? , a unsupervised algorithm

Justin-Le commented 7 years ago

In hyperspectral data, the pixel is the data point, not the image. One image contains thousands to millions of data points. For example, the one that I used (Pavia) contains about 1 million points, which is not huge but still big enough for learning.

lotfiGipsyKing commented 7 years ago

can you explain to me the global architecture that you used, did you use a dimension réduction algorithm, spatial and spectral features extraction méthodes, Morphogical algorithm ????

Justin-Le commented 7 years ago

Details can be found in the paper, "A Deep Belief Network for Classifying Remotely-Sensed Hyperspectral Data" in ISVC 2015.