upm Namespace Reference

Classes
class	BAD

Functions
static void	rectifyBoxes (const std::vector< BAD::BoxPairParams > &boxes_params, std::vector< BAD::BoxPairParams > &out_params, const cv::KeyPoint &kp, float scale_factor=1, const cv::Size &patch_size=cv::Size(32, 32))
	Rectifies the coordinates of the box pairs measurement functions with the keypoint location parameters. More...
static float	computeBadResponse (const BAD::BoxPairParams &box_params, const cv::Mat &integral_img)
	Computes the Box Average Difference, measuring the difference of gray level in the two square regions. More...
static bool	isKeypointInTheBorder (const cv::KeyPoint &kp, const cv::Size &img_size, const cv::Size &patch_size={32, 32}, float scale_factor=1)
	Function that determines if a keypoint is close to the image border. More...

Detailed Description

Copyright

2021 Xoan Iago Suarez Canosa. All rights reserved. Constact: iago..nosp@m.suar.nosp@m.ez@th.nosp@m.egra.nosp@m.ffter.nosp@m..com Software developed in the PhD: Low-level vision for resource-limited devices Repo link: https://github.com/iago-suarez/efficient-descriptors License: Apache 2 TODO(rsoussan): Remove this and BAD.cpp and switch to opencv TEBLID if opencv version upgraded.

Function Documentation

computeBadResponse()

static float upm::computeBadResponse ( const BAD::BoxPairParams &  box_params, const cv::Mat &  integral_img  )

inlinestatic

Computes the Box Average Difference, measuring the difference of gray level in the two square regions.

Parameters

box_params	The weak-learner parameter defining the size and locations of each box.
integral_img	The integral image used to compute the average gray value in the square regions.

Returns

The difference of gray level in the two squares defined by box_params

isKeypointInTheBorder()

static bool upm::isKeypointInTheBorder ( const cv::KeyPoint &  kp, const cv::Size &  img_size, const cv::Size &  patch_size = {32, 32}, float  scale_factor = 1  )

inlinestatic

Function that determines if a keypoint is close to the image border.

Parameters

kp	The detected keypoint
img_size	The size of the image
patch_size	The size of the normalized patch where the measurement functions were learnt.
scale_factor	A scale factor that magnifies the measurement functions w.r.t. the keypoint.

Returns

true if the keypoint is in the border, false otherwise

rectifyBoxes()

static void upm::rectifyBoxes ( const std::vector< BAD::BoxPairParams > &  boxes_params, std::vector< BAD::BoxPairParams > &  out_params, const cv::KeyPoint &  kp, float  scale_factor = 1, const cv::Size &  patch_size = cv::Size(32, 32)  )

inlinestatic

Rectifies the coordinates of the box pairs measurement functions with the keypoint location parameters.

Parameters

boxes_params	The input parameters defining the location of each pair of boxes inside the normalized (32x32) patch.
out_params	The output parameters, now the boxes are located in the full image.
kp	The keypoint defining the offset, rotation and scale to be applied
scale_factor	A scale factor that magnifies the measurement functions w.r.t. the keypoint.
patch_size	The size of the normalized patch where the measurement functions were learnt.