0
$\begingroup$

I am currently working on a problem for which the topographic data is in very different resolution. Let say I have data of 20x20 with 1km2 tiles and also high resolution data of 50m2 tiles. I would like to combine both for input in a CNN. To make things more spicy I don't care about the 50m2 when it is far away from the center, that is why I would like to use an 'image' multi resolution, aka resolution low in the edges but higher in the center. That would be like human vision, only high detailed in the center... Then I would combine that multi-resolution image with my 1km2 data

Do you know any research done on such a CNN ?

Only found that one for now Multi-Resolution Feature Fusion for Image Classification of Building Damages with Convolutional Neural Networks

Thank you for you help

| improve this question | |
$\endgroup$
  • $\begingroup$ So your input is an image divided into 20x20 tiles where some of them represent $1$ km$^2$ area and some of them represent $50$ m$^2$ area, and they are all mixed up ? What are you trying to classify ? Did you try "regular" architectures ? $\endgroup$ – Brale Apr 20 at 11:28
  • $\begingroup$ I have many layers of 1 km2 in my image of 20x20. I want to add layers from data given at 50 m2 resolution, but I want to represent the same 20km2. Hence the idea of multi-resolution layer.... Actually I do avg pooling to produce a lower (1km2) resolution layer. That being said, topographic at 1km2 resolution is useless for my application $\endgroup$ – Q. Fisch Apr 20 at 13:28
  • $\begingroup$ I think Multi-resolution so far I have not seen a multi resolution CNN, and in my opinion it hasn't been there yet. $\endgroup$ – Fajar Hidayat Apr 21 at 10:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.