With the rise of multimedia devices, now there are many images with different sizes and qualities that must be displayed with the use of different devices. Hence, users need to display an image on a device with the resolution greater or less than of the image’s resolution. Therefore, to display an image it is probable to resize the image before displaying it. To resize the images there are two general methods containing traditional methods and content-aware methods. Traditional methods are simple methods that remove, insert or modify pixels using a general rule without any consideration to the image’s content. Hence, quality of these methods is not very good. In opposite, content-aware methods consider image’s content and preserve pixels that are more notable for human visual system. Hence, quality of these methods is better than traditional methods. Seam carving is one of these content-aware methods that is simple and efficient for many images. But whereas this method is naturally a sequential process, it is computationally hard and is a slow method. In this thesis we propose two new approaches to accelerate original method. The first approach speeds-up the method by 5.5 times in average and the second one accelerate it by 13.8 in average. We also propose a new energy map that it use to preserve image content more efficient. Keywords : Seam, content-aware resizing, pixel energy
