HOG (Histogram of Oriented Gradients) MEX Implementation

Rectangular histogram of gradients (R-HoG) descriptor MEX implementation.

Description

Descriptor implemented according to Histograms of Oriented Gradients for Human Detection [Dalal, Triggs, 2005], with added variance normalization for the input image.

Compile using mex hog.cxx.

All the code is in the header file, hog.h, so it is possible to include that header file into your cpp application and use it directly. hog.cxx simply serves as an entry point for matlab.

See the comments in hog.cxx for usage.

• Required Products: MATLAB
• MATLAB release: MATLAB 7.13 (R2011b)

[pushplata]

hog.cxx
d:\new folder\data\programz\hog (3)\hog.h(46) : error C2057: expected constant expression
d:\new folder\data\programz\hog (3)\hog.h(46) : error C2466: cannot allocate an array of constant size 0
d:\new folder\data\programz\hog (3)\hog.h(46) : error C2133: 'gauss' : unknown size
d:\new folder\data\programz\hog (3)\hog.h(52) : error C2065: 'M_PI' : undeclared identifier
d:\new folder\data\programz\hog (3)\hog.h(77) : error C2065: 'M_PI' : undeclared identifier
hog.cxx(103) : warning C4800: 'int' : forcing value to bool 'true' or 'false' (performance warning)

C:\PROGRA~1\MATLAB\R2012A\BIN\MEX.PL: Error: Compile of 'hog.cxx' failed.

following errors occurred after compilation.