Video target area shielding method and device

Abstract

The invention discloses a video target area shielding method and a video target area shielding device. The video target area shielding method comprises the steps of: determining a target area and a target feature in the target area in a video image through analyzing the video image; tracking the video image according to the target feature, and determining a position of the target feature in the video image; performing machine learning on the target area based on the target feature, and determining an area replacement image matched with the target area in a preconfigured area replacement image database; and placing the area replacement image in the target area by adopting a seamless splicing technique, so that the area replacement image and a background image of the target area are spliced in a seamless manner. According to the video target area shielding method and the video target area shielding device, the target area is determined through analyzing and tracking the video image, the area replacement image matched with the target area is determined by adopting a machine learning method, and the area replacement image is placed in the target area by adopting the seamless splicing technique, so that the processed image is connected and fused with the background image of the video without affecting the esthetic appearance.

Description

technical field [0001] The present invention relates to the field of video images, in particular to a method and device for blocking a video target area. Background technique [0002] At present, in video images, in order to protect personal privacy, people usually wear hats or sunglasses to avoid revealing personal identities, but this method will bring inconvenience to the parties, so in order to reduce the burden on the parties, the In the prior art, the occlusion of the person images in the video images is generally achieved by occluding or mosaicing the video images, so as to achieve the purpose of protecting personal privacy. [0003] However, for the above two video image processing methods in the prior art, there are some problems that need to be solved. First of all, as far as the method of blocking the video image in the prior art is concerned, it is actually through The position of the original image directly covers a replacement image, and the replacement image ...

AI Technical Summary

Problems solved by technology

[0003] However, for the above two video image processing methods in the prior art, there are some problems that need to be solved. First of all, as far as the method of blocking the video image in the prior art is concerned, it is actually through The position of the original image directly covers a replacement image, and the replacement image only achieves the effect of occluding the original image. The replacement image cannot be connected and integrated with the image scene in the video. For example, it is a An existing picture is pasted directly to the position of the face, but the picture cannot match the face, nor can it be connected and fused with the surrounding background image

[0004] On the other hand, as far as the method of mosaicing a video image in the prior art is concerned, it locks the position of the human eye in the video, and then zooms in and out the fixed rectangular area where the human eye is located, so as to achieve The effect of mosaicing by human eyes, although this method protects personal privacy to a certain extent, this method of mosaicing video images also has the problem of visually affecting the appearance

[0005] For the above-mentioned problems in related technologies, no effective solution has been proposed yet

Images