Automatic image cropping using saliency map

Cropping is one of the image processing techniques mainly utilized to remove unwanted areas from an image and to improve presentation of visual elements in an image. Selection of the picture window via processing is subjective in nature and typically a manual task. This paper describes an approach for automatic image cropping which is based on saliency map detection to preserve salient region in image. Proposed method uses resultant image of saliency map detection to eliminate unwanted regions from an image. Further, entire process is repeated iteratively until optimum cropping is achieved. Here, cropped image improves visual composition of image and it used in a variety of application in printing industry, photography, and cinematography. Experimental results using subjective analysis show effectiveness of the proposed method.