Picture processing method and apparatus, device, and storage medium

Patent metadata

Field	Value
Publication number	US-12136156-B2
Application number	US-202218090181-A
Country	US
Kind code	B2
Filing date	Dec 28, 2022
Priority date	Aug 19, 2020
Publication date	Nov 5, 2024
Grant date	Nov 5, 2024

Abstract

A picture processing method and apparatus, a device, and a storage medium are provided. The method comprises: obtaining pictures to be processed, the pictures to be processed comprising a target object segmentation picture and a background completion picture from a same original picture; respectively performing effect processing on the target object segmentation picture and the background completion picture on the basis of a picture type of the target object segmentation picture to obtain respective effect pictures; and synthesizing the respective effect pictures into a rotation animation corresponding to the original picture.

First claim

The invention claimed is:

1. A picture processing method, comprising:
obtaining pictures to be processed, the pictures to be processed comprising a target object segmentation picture and a background completion picture from a same original picture;
performing effect processing on the target object segmentation picture and the background completion picture separately, according to a picture type of the target object segmentation picture, to obtain an effect picture corresponding to the target object segmentation picture and an effect picture corresponding to the background completion picture, comprising:
identifying the picture type of the target object segmentation picture,
determining an effect processing scheme corresponding to the identified picture type of the target object segmentation picture, wherein the picture type of the target object segmentation picture is a full body picture or a half body picture, and wherein the full body picture and the half body picture correspond to effect processing schemes different from each other, and
performing the effect processing on the target object segmentation picture and the background completion picture separately with the determined effect processing scheme;
synthesizing the effect picture corresponding to the target object segmentation picture and the effect picture corresponding to the background completion picture to generate an effect video corresponding to the original picture; and
generating an album from the plurality of effect videos corresponding to the plurality of the original pictures.

2. The method according to claim 1 , wherein the performing effect processing on the target object segmentation picture and the background completion picture separately, according to a picture type of the target object segmentation picture, to obtain an effect picture corresponding to the target object segmentation picture and an effect picture corresponding to the background completion picture comprises:
if the picture type of the target object segmentation picture is the full body picture,
translating and enlarging the target object segmentation picture according to a preset first direction to obtain the effect picture corresponding to the target object segmentation picture, the preset first direction being rightward or leftward; and
rotating a perspective view of the background completion picture according to a preset second direction to obtain the effect picture corresponding to the background completion picture, the preset second direction being leftward or rightward, and the preset first direction being different from the preset second direction.

3. The method according to claim 2 , wherein the synthesizing the effect picture corresponding to the target object segmentation picture and the effect picture corresponding to the background completion picture to generate the effect video corresponding to the original picture comprises:
synthesizing the effect picture corresponding to the target object segmentation picture and the effect picture corresponding to the background completion picture to generate a rotating animation corresponding to the original picture.

4. The method according to claim 3 , wherein the synthesizing the effect picture corresponding to the target object segmentation picture and the effect picture corresponding to the background completion picture to generate the rotating animation corresponding to the original picture comprises:
if the picture type of the target object segmentation picture is the half body picture,
determining a connection point between a target object in the effect picture corresponding to the target object segmentation picture and the effect picture corresponding to the background completion picture; and
synthesizing the effect picture corresponding to the target object segmentation picture and the effect picture corresponding to the background completion picture to generate the rotating animation corresponding to the original picture based on the connection point.

5. The method according to claim 1 , wherein the performing effect processing on the target object segmentation picture and the background completion picture separately, according to a picture type of the target object segmentation picture, to obtain an effect picture corresponding to the target object segmentation picture and an effect picture corresponding to the background completion picture comprises:
if the picture type of the target object segmentation picture is the half body picture,
enlarging the target object segmentation picture according to a forward direction to obtain the effect picture corresponding to the target object segmentation picture; and
rotating a perspective view of the background completion picture for an effect of a front leaning to a preset angle, to obtain the effect picture corresponding to the background completion picture.

6. The method according to claim 1 , wherein the target object segmentation picture comprises a portrait segmentation picture.

7. The method according to claim 1 , wherein before the performing effect processing on the target object segmentation picture and the background completion picture separately, according to a picture type of the target object segmentation picture, to obtain an effect picture corresponding to the target object segmentation picture and an effect picture corresponding to the background completion picture, the method further comprises:
determining whether a target object in the target object segmentation picture comprises a half body object;
determining that the picture type of the target object segmentation picture is the half body picture if the target object in the target object segmentation picture comprises a half body object; and
determining that the picture type of the target object segmentation picture is the full body picture if the target object in the target object segmentation picture does not comprise a half body object.

8. The method according to claim 1 , further comprising:
extracting a target object from the original picture to obtain the target picture segmentation picture;
understanding content of the original picture; and
performing padding, based on the content of the original picture, in an area of the original picture where the target object is removed, to obtain a complete picture as the background completion picture.

9. A non-transitory computer-readable storage medium having instructions stored thereon, wherein the instructions, when being executed on a terminal device, cause the terminal device to implement:
obtaining pictures to be processed, the pictures to be processed comprising a target object segmentation picture and a background completion picture from a same original picture;
performing effect processing on the target object segmentation picture and the background completion picture separately, according to a picture type of the target object segmentation picture, to obtain an effect picture corresponding to the target object segmentation picture and an effect picture corresponding to the background completion picture, comprising:
identifying the picture type of the target object segmentation picture,
determining an effect processing scheme corresponding to the identified picture type of the target object segmentation picture, wherein the picture type of the target object segmentation picture is a full body picture or a half body picture, and wherein the full body picture and the half body picture correspond to effect processing schemes different from each other, and
performing the effect processing on the target object segmentation picture and the background completion picture separately with the determined effect processing scheme;
synthesizing the effect

Assignees

• Beijing Bytedance Network Tech Co Ltd

Inventors

• Zhang Shupeng
• Li Yameng
• Wang Yitong
• Liu Sixue
• Tian Yuan
• Liao Jianjun
• Wen Jie

Classifications

• Y02D10/00

  Energy efficient computing, e.g. low power processors, power management or thermal management · CPC title

• G06T2210/32

  Image data format · CPC title

• G06T2207/20221

  Image fusion; Image merging · CPC title

• G06T11/60

  Creating or editing images; Combining images with text · CPC title

• G06T3/60

  Rotation of whole images or parts thereof · CPC title