Multiple Neural Network Based Object Segmentation in a Sequence of Color Image Frames

摘要

An image processing apparatus and method for object segmentation in color image frames are provided. The image processing apparatus generates, based on the first neural network model, a first foreground mask and a second foreground mask for the object of interest in a first color image frame and an upcoming color image frame of the sequence of color image frames, respectively. The image processing apparatus determines a third foreground mask based on the interpolation of the first foreground mask and the second foreground mask, and updates the third foreground mask to the fourth foreground mask by using the second neural network model. The image processing apparatus divides the object of interest from at least the first color image frame, the upcoming color image frame, and the intermediate color image frame by using the first foreground mask, the second foreground mask, and the fourth foreground mask, respectively.