학술논문
휘도 인식 모듈을 통한 다중 노출 오류 보정 네트워크 / Multi Exposure Correction Network via Luminance Aware Module
Document Type
Dissertation/ Thesis
Author
Source
Subject
Language
Korean
Abstract
이미지 촬영 중 환경 문제로 인한 노출 오류는 이미지 정보 손실을 유발한다. 노출 오류를 수정하기 위해 많은 알고리즘이 제안되었지만, 과노출/저노출의 다중 노출 오류 영상을 수정하기 위한 기존 방법을 해결하기는 어렵다. 본 논문에서는 휘도 인식 모듈을 통한 다중 노출 보정 네트워크를 제안한다. 휘도 인식 모듈은 영상의 밝기 비교를 통해 영상의 전체적인 휘도 및 노출 오차를 조절한다. 제안된 네트워크는 다중 노출 오류 수정 작업에서 기존 방법보다 높은 PSNR 및 SSIM 점수를 기록했다.
Exposure error caused by an environmental problem during photographing of an image causes loss of image information. Many algorithms have been proposed to correct exposure error, but it is difficult to solve the existing method for correcting multiple exposure error images of over-exposure/low exposure. In this paper, we propose multiple exposure correction network via luminance aware module. The luminance aware module adjusts the overall luminance and exposure error of the image through the brightness comparison of images. The proposed network recorded higher PSNR and SSIM scores than conventional methods in multiple exposure error correction tasks.
Exposure error caused by an environmental problem during photographing of an image causes loss of image information. Many algorithms have been proposed to correct exposure error, but it is difficult to solve the existing method for correcting multiple exposure error images of over-exposure/low exposure. In this paper, we propose multiple exposure correction network via luminance aware module. The luminance aware module adjusts the overall luminance and exposure error of the image through the brightness comparison of images. The proposed network recorded higher PSNR and SSIM scores than conventional methods in multiple exposure error correction tasks.