형태학적 마스크 확장을 이용한 웨이블릿 계수의 영상압축 알고리즘

Abstract

본 논문에서는 EZW에서 사용한 제로트리 개념을 중요계수에 적용시키고, 선택된 중요계수를 중심으로 형태학적 확장을 시행하여 중요계수의 위치를 코딩하였다. 형태학적 확장을 적용시키는 과정에서 마스크의 중심을 중요계수에만 한정시킨다면, 어떤 경우에는 단독 중요계수(isolated significant coefficients)를 코딩하지 못하는 경우가 발생하게 되므로 본 논문에서는 마스크의 중심이 적응적으로 변화하도록 하여 단독 중요계수의 손실이 최소가 되도록 하였다. 적응적 마스크의 확장을 이용한 코딩법은 불필요한 코딩의 수를 줄이게 되고, 비트율(bit-rate)을 낮출 수 있음을 실험을 통하여 확인할 수 있었다.

In this paper, 'the location of the significant coefficients is possibly found with the mask operations and the magnitude of the coefficients is reconstructed using SAQ by taking advantage of EZW and MRWD. If only significant coefficients are used on the center of the mask operations, the isolated significant coefficients are not able to be coded in some cases, therefore the center of mask is changed adaptively. The results of experiments show that the number of coding and bit-rate is lowered by the method of coding with the adaptive dilation of the mask.

In this paper, 'the location of the significant coefficients is possibly found with the mask operations and the magnitude of the coefficients is reconstructed using SAQ by taking advantage of EZW and MRWD. If only significant coefficients are used on the center of the mask operations, the isolated significant coefficients are not able to be coded in some cases, therefore the center of mask is changed adaptively. The results of experiments show that the number of coding and bit-rate is lowered by the method of coding with the adaptive dilation of the mask.