A Fast Seam Carving Method Based on Merging Seams in Subimages

Display devices based on its different screens andresolutions, need image resizing to retain the image’s quality.Common image resizing methods cannot save or protectimportant objects or its results are non-photorealistic. Seamcarving as a new method has been widely used for contentawareimage and video resizing with little distortion incomparison with common methods.Unfortunately, seam carving is a complex algorithm and forhigh resolution videos or images has a long run time and is notusable in real-time applications. In this paper, a novel fastmethod in order to accelerate simple seam carving and decreasecomputational burden is presented. In this method image isdivided into three equal horizontal or vertical sections, while thetraditional seam carving is applied to the middle section. In thetop and down sections, the algorithm estimates seam with respectto the middle part seam using an approximated Dijkstra method.Experiments have demonstrated better computational efficiencyof presented method when it faces the current seam carvingmethod .It is also preserving the image’s information aseffectively as the original seam carving method.

Keywords:

Seam carving , image resizing , Dijkstra method