Displaying images on the web is always a case of walking the fine line between optimizing the image for the web and making sure the image doesn't lose its clarity or its intended purpose. There are two ways of resizing an image which is popularly used by web masters and lay persons alike. They are cropping an image and scaling an image to the desired size. The former method will most possibly end with the image losing its fine details where as using the latter method will reduce the image's clarity.
Now it seems a new method of resizing an image has been demonstrated by two Israeli professors, Shai Avidan and Ariel Shamir which ensures that the image retains both its functionality as well as its clarity. This method is called re-targeting. The idea is to algorithmically determine which areas of the image are important and then delete other areas retaining those which are important.
Fig: Re-targeted vs cropped vs scaled imageAt present none of the graphics suites including Photoshop and Gimp provide tools to re-target an image. Perhaps if this new algorithm (or method) is released under an open license, we could see this functionality being incorporated in Gimp in the future. You can read more about this new method here (PDF document). It is popularly known as 'Seam Carving for Content-Aware Image Resizing'.
For now watch this Youtube video which explains this interesting process in more detail.
No comments:
Post a Comment