Image character removing method, system and device based on gate circulation unit

Abstract

The invention belongs to the field of computer graphics and computer vision, particularly relates to an image character removing method, system and device based on a gate circulation unit, and aims tosolve the problem that characters in an image cannot be effectively removed due to the fact that accurate character strokes cannot be extracted in the prior art. The method comprises the following steps: constructing a stroke-level binary mask generation model based on one or more inverse residual modules connected in series and a gate cycle unit, and obtaining a stroke-level binary mask of an input image based on the input image and a character region mask; and constructing a character removal model based on the one or more encoders, the one or more decoders and the feature processing module, and obtaining a character-removed image based on the fusion features of the input image, the character region mask and the stroke-level binary mask. The fine stroke-level character mask can be obtained, so that the model is effectively guided to remove the characters with high quality, and the rendering and backfilling effects in the character removing process are improved.

Description

technical field [0001] The invention belongs to the field of computer graphics and computer vision, and in particular relates to a method, system and device for removing image characters based on a gate cycle unit. Background technique [0002] With the development and progress of society, images are ubiquitous as an important form of information transmission. When it transmits valuable information, it will also cause leakage of private information. For example, when taking photos or collecting data, some personal information (including ID number, phone number, etc.) will inevitably appear in the image. Image text removal technology can protect privacy by removing sensitive information in images. At the same time, this technology is also widely used in image and video editing, text translation and other tasks. [0003] Early image text removal methods were mainly based on color histograms or thresholding. Khodadadi et al. estimate the background and text colors in the can...

AI Technical Summary

Problems solved by technology

[0016] In order to solve the above-mentioned problems in the prior art, that is, the prior art cannot extract accurate character strokes, and thus cannot effectively remove the characters in the image, the present invention provides a method for removing characters in images based on gated cycle units, The method includes the following steps:

Images