A Method of Portrait Background Cutout

Abstract

The invention discloses a portrait background cutout method. The portrait background cutout method comprises the following steps of acquiring the jaw position of a portrait in a picture, marking foreground pixel, background pixel, possible foreground pixel and possible background pixel of an area above the jaw, marking foreground pixel, possible foreground pixel and possible background pixel of an area under the jaw, segmenting the photo according to pixel marks on the whole photo, automatically generating an initial Trimap picture along segmented boundaries, finding hair belonging to foreground in the background area of the Trimap picture, detecting hair gaps belonging to background in the hair area, automatically generating the final Trimap picture according to the hair area and the hair gaps, calculating according to the final Trimap picture to obtain an alpha picture and combining the alpha picture with new background to obtain a new composite result picture the portrait background cutout method can automatically process pictures in batch and enables the transition between background and foreground in the final composite result picture to be smooth and real.

Description

technical field [0001] The invention relates to an image cutout method, in particular to a portrait background cutout method. It belongs to the field of image processing. Background technique [0002] As we all know, the research on image matting has a history of more than 20 years. The so-called matting is to accurately extract the foreground of an object from the background, which can be expressed by the following formula: [0003] I z = α z f z +(1-α z )B z [0004] Among them, α z is the transparency of pixel z: when α z = 0, the pixel z is the background, when α z = 1, pixel z is foreground; F z is the foreground color value of pixel z, B z is the background color value of pixel z, I z The foreground color value is mixed with the background color value to obtain the color value of the pixel z itself. Although most pixels are determined foreground or background, it is necessary to accurately estimate α in the transition region between foreground and backgrou...

AI Technical Summary

Problems solved by technology

Therefore, all of these methods have the following problems: 1) human interaction is required to specify all or part of the foreground and background regions; 2) due to the complexity of the natural image background, small errors in the foreground boundary matting can easily be synthesized Especially for ID photos, because the composite background of ID photos is solid color, tiny matting defects are easily enlarged in the composite image, so there is a higher precision requirement for the foreground boundary

[0012] 1) It relies heavily on Trimap images or scribbles that are manually input interactively as input. It is difficult to fully automate the entire matting process, and the application efficiency is low when a large number of photos need to be processed frequently.

[0013] 2) The boundary between the foreground and the background is blurred, blurred or excessively unsmooth

Images