Combination of saliency histogram equalisation and seam carving for image resizing

Compressed domain image resizing based on saliency histogram equalisation and seam carving

“…This paper takes the SH approach to the problem of CAIR in the compressed domain. Two of the related algorithms, SHE [11] and SHE‐SC [12] are reviewed in this section. The advantage of SHE and SHE‐SC is due largely to computational simplicity.…”

“…Thus, is inversely proportional to , and the least significant seam (LSS) of quads is the one with the maximum sum of quad sizes. In other words, the LSS of quads in the vertical direction is as follows: which can be efficiently obtained via dynamic programming [12]. By adjusting the width of each quad in appropriately, the width of M can be resized accordingly.…”

Video retargeting based on SH equalisation and seam carving

“…Pixels with great saliency values are often in the ROI, which need to be preserved as much as possible; on the other hand, those with small saliency values in the region of non‐interest (RON) can be ignored in the retargeted images/videos. This section reviews the SC algorithm [2] that is representative of the discrete approach, the SHE algorithm [13] that is classified into the continuous approach, and the improved SHE algorithm by combining SHE with SC [18].…”

“…In the SHE algorithm, the input mesh that represents the original image is obtained by equalising its SH, and the output mesh that represents the retargeted image is simply a uniform mesh. To improve the performance of SHE, a hybrid algorithm that combines SHE with SC has been proposed recently [18]. Specifically, let be the non‐uniform mesh obtained by SHE; and V , E , and Q are sets of vertices, edges, and quads, respectively.…”

“…[15, 16], warp a uniform input mesh representing original frames into a non‐uniform output mesh representing retargeted frames. Nevertheless, both input and output meshes can be non‐uniform for the sake of flexibility [18]. Krähenbühl et al improved the visual quality of retargeted videos through the use of high‐level interactions [19].…”