Feature extractions to optimize scanned images

A non-transitory machine-readable storage medium including instructions that when executed cause a processing resource to:

• apply a feature extraction process to extract a gamut-based feature included in a plurality of pixels of a scanned image by:
• converting a color space of the plurality of pixels from an input color space to an output color space;
• segmenting the output color space into a plurality of hue segments each including a subset of pixels of the plurality of pixels;
• and
• determining which of the plurality of converted and segmented pixels are associated with a highlighter marking of the scanned image;
• apply a classification model to the scanned image;
• and
• optimize the scanned image based on the classification of the scanned image.

A method, comprising:

• applying, by an imaging device, a feature extraction process including:
• converting a color space of a plurality of pixels of a scanned image from an input color space to a polar coordinate representation of an opponent color space;
• segmenting the polar coordinate representation of the opponent color space into a plurality of hue segments each including a subset of pixels of the plurality of pixels;
• comparing each pixel of the plurality of pixels in each hue segment of the plurality of hue segments with a corresponding hue segment included in predefined imaging device gamut segments to determine a distance of each pixel from a boundary of each respective predefined imaging device gamut segment;
• and
• determining whether the distances of each pixel from the boundary of each respective predefined imaging device gamut segment satisfy constraints to determine which of the plurality of pixels are highlighter pixels;
• applying a classification model to the scanned image;
• and
• optimizing the scanned image based on the classification of the scanned image for a print job.