Create a new or load an existing HOG descriptor and detector

hog = cv.HOGDescriptor()
hog = cv.HOGDescriptor('PropertyName', propertyValue, ...)
hog = cv.HOGDescriptor(filename)

Input

• filename Filename of existing HOG descriptor config to load from (XML or YAML). This is handled by the load method.

Output

• hog HOG descriptor object.

Options

• WinSize Detection window size. Align to block size and block stride. default [64,128]
• BlockSize Block size in pixels. Align to cell size. default [16,16]
• BlockStride Block stride. It must be a multiple of cell size. default [8,8]
• CellSize Cell size. default [8,8]
• NBins Number of bins in the calculation of histogram of gradients. default 9
• DerivAperture Derivative of aperture. default 1
• WinSigma Gaussian smoothing window parameter. default -1 (corresponds to sum(BlockSize)/8)
• HistogramNormType Histogram normalization method. default 'L2Hys'
• L2HysThreshold L2-Hys normalization method shrinkage. default 0.2
• GammaCorrection Flag to specify whether the gamma correction preprocessing is required or not. default true
• NLevels Maximum number of detection window increases. default 64
• SignedGradient Flag to specify whether orientations range in 0-180 (false) or 0-360 (true) degrees. 0-180 mean that positive/negative directions count as the same histogram bin. default false