ClusterNet: Detecting Small Objects in Large Scenes by Exploiting Spatio-Temporal Information

[@mertkayalk3074]

Is it a visual image (.jpeg) used in the video, or an orthophoto (.tiff) produced using photogrammetry, or a satellite image (.img)? If this is a simple image, is it possible to do the same for orthophotos produced using multiple high resolution photos?

[@erikschiegg68]

So it stays at good old tiling for _non_ moving small objects in a large scene?

[@TerboBlitz]

Interesting. Can you provide links to published references?

[@karankatiyar5414]

@@RodneyLaLonde githib?

[@RodneyLaLonde]

@@karankatiyar5414 unfortunately this research was funded by Lockheed Martin and I cannot release the code publicly for intellectual property reasons.

[@karankatiyar5414]

@@RodneyLaLonde are there any other networks you can suggest ? I am trying to detect object at 80 m height taken by UAV and I want an algo that can detect a certain object with a little bit of variation in size

[@RodneyLaLonde]

@@karankatiyar5414 A set of strong works in that area is the two (CVPR and PAMI) by Rozantsev et al. you can find at this link (cvlab.epfl.ch/research/research-unmanned/research-unmanned-detection/). We are also working on such problems here at the CRCV. You can also try such networks as Faster R-CNN and the like. Another strong researcher in this area is my friend Michael Teutsch, I would recommend looking at his publications.

[@kudzo3920]

@@RodneyLaLonde Hi, Sir. Please This link is not working any help would be appreciated.