#include <psi_likelihood.h>
- Parameters
-
data | Bounding box set in format: data[frame][actor]. These should be pre-thinned, so the time delta between data boxes equals the time step between simulator outputs (TODO need a better way to explain this). |
cam | A "standardize" perspective camera (screen origin at center of image, y-axis pointing upward). |
width | Width of the video frame |
height | height of the video frame |
- Note
- The camera provided MUST be defined using "standard" screen coordinates; i.e. the origin of screen coordinates is located in the center of the frame, the x-axis points right, y-axis points up, and the z-axis points out of the screen.
- Parameters
-
p_frame_count | probability of each extra frame |
std::vector<std::pair<double, Bbox> > kjb::psi::Cylinder_world_likelihood::blocks_to_boxes |
( |
const std::vector< Cuboid > & |
blocks | ) |
const |
convert 3d blocks into 2d bounding boxes, with depths from camera
std::vector< std::pair< double, Bbox > > kjb::psi::Cylinder_world_likelihood::entities_to_boxes |
( |
const std::vector< std::vector< Entity_state > > & |
entities | ) |
const |
convert 3d entities into 2d bounding boxes, with depths from camera
Take a set of entity states for N humans and convert them into N into bounding boxes, sorted descended by depth from camera.
double kjb::psi::Cylinder_world_likelihood::operator() |
( |
const Model & |
m | ) |
const |
|
virtual |
The documentation for this class was generated from the following files: