@article{478,
  author = {Claudio S. V. C. Cavalcanti, Herman Martins Gomes, José Eustáquio Rangel de Queiroz},
  title = {An Extended Evaluation of Methods for Portrait Cropping},
  journal = {Journal of Multimedia Processing and Technologies},
  year = {2010},
  volume = {1},
  number = {4},
  doi = {},
  url = {http://www.dline.info/jmpt/fulltext/v1n4/3.pdf},
  abstract = {This paper is an extension of a previous work by the same authors [1] that proposed and evaluated an automatic method to guide automatic portrait cropping for different aspect ratios. The development of automatic methods for image cropping in general are motivated by the large variety of aspect ratios used for both image rendering (e.g. print media, TV, cinema screens etc.) and image acquisition devices (e.g. still and video cameras, scanners etc.). In order to maintain the image’s original aspect ratio when adjusting for a different media, some level of cropping may be required, but the Automatic Zoom & Crop method (AZ&C) may not produce satisfactory results regarding image contents. Our method analyses images and estimates the relevant content areas, avoiding distortions and main subject chopping. The analysis is performed by four feature extractors, each producing a grayscale image which indicates relevant image areas. The outputs of these extractors are then combined by means of a Genetic Algorithm (GA) optimization. Experiments involving a subjective evaluation of a set of automatically cropped images have shown that 77% of the 35 participants considered images adjusted by the Multi Saliency Cropping Method (MSCM) better or similar than the outputs of the AZ&C method. In order to further investigate the combination of the feature extractors in the MSCM method, an extended evaluation has been performed, which involved the use of a database of 100 images for ranking method variations using combined and isolated features in addition to the AZ&C method. The results have shown that in 80% of the participants opinions the MSCM method is considered equal or superior to the variations using isolated features and the AZ&C method.},
}