Hello Mark,
Normalizing your images is an important step that can affect the quality of the extracted features. Normalizing the grayscale values ensures that the input to the HOG algorithm is consistent across different images. It also helps mitigate the effects of contrast and illumination differences.
Moreover, HOG is sensitive to edges and gradients, and normalization can help ensure that these features are not overshadowed by variations in lighting.
Using “rgb2gray”, you can convert an RGB image to grayscale, but it does not perform normalization.
