Object segmentation in a sequence of color image frames based on adaptive foreground mask upsampling

A method, comprising:

• in an image processing apparatus comprising a first type of sensor, a second type of sensor, and a control circuitry:
• capturing, by the first type of sensor, a sequence of color image frames comprising a current color image frame and a previous current color image frame;
• receiving, by the control circuitry, a first foreground mask for an object-of-interest in the previous color image frame;
• estimating, by the control circuitry, a second foreground mask for the object-of-interest in the current color image frame at a first image resolution, based on the first foreground mask, and at least a shape of a portion of a human body in the current color image frame;
• upsampling, by the control circuitry, the estimated second foreground mask of the current color image frame from the first image resolution to a second image resolution, wherein the upsampled second foreground mask comprises at least one pixel with an unknown mask value;
• selecting, by the control circuitry, at least one region-of-interest in the current color image frame, based on a spatial position of the at least the one pixel of the upsampled second foreground mask;
• and
• upscaling, by the control circuitry, the upsampled second foreground mask by an expectation parameter generated based on a kernel density-based estimation for the at least one pixel in the selected at least one region-of-interest in the current color image frame, wherein the upscaled second Foreground mask
• exhibits a minimum foreground mask error that is caused by upsampling with respect to the estimated second foreground mask.