Machine learning feature vector generator using depth image foreground attributes

An apparatus, comprising:

• at least one processor;
• and
• at least one memory coupled to the at least one processor, the at least one memory comprising instructions that, when executed by the at least one processor, cause the at least one processor to:
• access a depth image comprising a plurality of image pixels;
• access an image segmentation mask comprising a plurality of segmentation pixels that correspond to at least a portion of the plurality of image pixels, at least a portion of the plurality of segmentation pixels comprising a plurality of foreground pixels representing at least one foreground object in the depth image;
• and
• for at least one seed pixel in a sequence of the plurality of image pixels:
• determine a segmentation pixel that corresponds to the at least one seed pixel,
• determine whether the segmentation pixel is one of the plurality of foreground pixels, and
• add feature data of the at least one seed pixel to a feature vector responsive to the segmentation pixel being one of the plurality of foreground pixels.