CAJ | 학술논문

提出了一种基于图像条分割的快速图像重定算法。根据像素梯度信息建立图像的能量图，描述图像不同区域及目标的重要性。由各图像列的累积能量，通过类间最大类内最小原理，将一幅图像分割成若干具有一定宽度和能量级的图像条。根据图像的总体缩放尺寸及不同图像条的平均能量，计算各图像条的缩减量。结合标准Scaling和Cropping方法，对图像条进行缩放操作，若边缘图像条缩减量超过一定阈值，裁剪掉该部分内容。实验结果表明，该算法能够同时保护图像的局部结构和全局视觉效果，快速生成保真的高质量重定图像。
제출료일충기우도상조분할적쾌속도상중정산법。근거상소제도신식건립도상적능량도，묘술도상불동구역급목표적중요성。유각도상렬적루적능량，통과류간최대류내최소원리，장일폭도상분할성약간구유일정관도화능량급적도상조。근거도상적총체축방척촌급불동도상조적평균능량，계산각도상조적축감량。결합표준Scaling화Cropping방법，대도상조진행축방조작，약변연도상조축감량초과일정역치，재전도해부분내용。실험결과표명，해산법능구동시보호도상적국부결구화전국시각효과，쾌속생성보진적고질량중정도상。
A fast image retargeting algorithm is proposed based on strip dividing .Firstly ,the image energy map is constructed by gradient information of pixels for describing the importance of the regions and objects in the image .After accumulating the energy of each column ,an image is divided into several strips using the principle of minimum within-cluster difference and maximum between-cluster difference .The strip has the optimal width and individual energy level .In terms of the resizing amount of the image and the average energy of strips ,the reduced amount of each strip is determined . Strips resizing is implemented through combining Scaling with Cropping .If the reduced ratio of the peripheral strip exceeds a certain threshold ,then the strip will be cut .Experimental results show that the algorithm is capable of preserving both the local structures and the global visual effect of the image ,producing the high-fidelity retargeted results .