Thank you for your effort and time. I tried everything from scratch. Took Theta V plugin, intergrated OpenCV 3.4.5 and wrote OpenCV functions in Java. Everything works like charm.
For Native static lib build, below link has good lead,
@Nitin_Bhaskar, thanks for all your work and sorry that the initial build of OpenCV didn’t work. I should have tested more fully prior to putting it up as the challenge template. Great to know that building from source yields a library that can work with the full set of OpenCV commands.
Update later in day
Tested @Nitin_Bhaskar’s sample code on both emulator and physical device RICOH THETA V. Both work great.
I was able to integrate OpenCV-4.0.1 android SDK with no-camera-opencv example. With this integration, earlier error with hog.detectMultiScale() was not seen. I have modified this example to apply bilateral filter on input image. Please check it out.
It works for me in a THETA AVD. I don’t see that the PROCESSED_IMAGE.PNG is changed, but the output in the emulator shows the image as having a blur (Bokeh effect?) applied to the background.
The original no-camera-opencv that nitin started with doesn’t save the processed file to storage. It’s not required for the challenge, but obviously the developer will want to add this feature in for the plug-in they submit for the contest (not the hardware challenge). For people that want to save the processed file to storage, the no-camera-template provides an example.