FilterMaskImageWithSize: Filter mask image with its size

[wkentaro]

Modified:

• jsk_perception/CMakeLists.txt
• jsk_perception/include/jsk_perception/multiply_mask_image.h
• jsk_perception/plugins/nodelet/libjsk_perception.xml Added:
• jsk_perception/cfg/FilterMaskImageWithSize.cfg
• jsk_perception/include/jsk_perception/filter_mask_image_with_size.h
• jsk_perception/src/filter_mask_image_with_size.cpp

[k-okada]

@wkentaro I have resolved conflict via web interface, so please confirm if that is ok.

[wkentaro]

Thanks. I rebased.

[knorth55]

I approved, but i don't see usecase of this nodelet. (Sample launch subscribe same mask image.)

[wkentaro]

I added flag ~use_reference to minimize the overhead caused by synchronizing.

[wkentaro]

@knorth55 Could u Please check the latest commit?

[wkentaro]

Please merge this.

[k-okada]

You can also use back projection method for detection motion area , if the camera is fixed on world frame-> ros-perception/opencv_apps#62
note that a camera with automatic white balance features does not work well with that algorithm, you'd better to use good camera. > @wkentaro

And if we can assume the moving object is human, we can use that for human detection application -> http://goldberg.berkeley.edu/pubs/acc-2012-visual-tracking-final.pdf > @kochigami @furushchev

This algorithm combines statistical background image estimation and per-pixel Bayesian segmentation. It was introduced by Andrew B. Godbehere, Akihiro Matsukawa, Ken Goldberg in their paper “Visual Tracking of Human Visitors under Variable-Lighting Conditions for a Responsive Audio Art Installation” in 2012. As per the paper, the system ran a successful interactive audio art installation called “Are We There Yet?” from March 31 - July 31 2011 at the Contemporary Jewish Museum in San Francisco, California.

[wkentaro]

You can also use back projection method for detection motion area , if the camera is fixed on world frame-> ros-perception/opencv_apps#62
note that a camera with automatic white balance features does not work well with that algorithm, you'd better to use good camera. > @wkentaro

Thank you for the info.

And if we can assume the moving object is human, we can use that for human detection application

I tried using fcn_object_segmentation.py for person detection for object data collection, but it was a bit difficult to segment the person and him/her holding objects.