Struct HogDescriptor 

pub struct HogDescriptor { /* private fields */ }

Histogram of Oriented Gradients descriptor.

Computes a feature descriptor based on gradient orientation histograms localized to fine-spatial regions (cells), grouped into blocks for contrast normalization.

Implementations

impl HogDescriptor

pub fn new(cell_size: usize, block_size: usize, nbins: usize) -> Self

Creates a new HOG descriptor.

• cell_size: Size of each cell in pixels (e.g., 8).
• block_size: Number of cells per block (e.g., 2).
• nbins: Number of orientation bins (e.g., 9).

pub fn compute<B: Backend>(&self, image: &Image<B>) -> Result<Tensor<B, 1>>

Computes the HOG descriptor for the given image.

The image is converted to grayscale, gradient magnitude and direction are computed, orientation histograms are built per cell, and blocks are L2-normalized. Returns a 1D tensor containing the concatenated block descriptors.