This nodelet is used to compute the saliency of an image.
The image can be composed of any feature space. However, the current nodelet takes only 3 channel BGR image and computes the saliency map.
The original code is implemented by the authors of the paper: Human detection using a mobile platform and novel features derived from a visual saliency mechanism.
Please cite this paper should you use the code.
However, there are minor changes from the original code including the MultiThreading support using OpenMP.
Currently only supports
BGR8encoding (3 channel RGB intensity image)
Output saliency map.
Number of threads used for execution
Prints the frame rate to the output image if true.
roslaunch jsk_perception sample_saliency_map_generator.launch