A Model of Saliency-Based Visual Attention for Rapid Scene Analysis
IEEE Transactions on Pattern Analysis and Machine Intelligence
Algorithms for Defining Visual Regions-of-Interest: Comparison with Eye Fixations
IEEE Transactions on Pattern Analysis and Machine Intelligence
Information Theoretic Measure for Visual Target Distinctness
IEEE Transactions on Pattern Analysis and Machine Intelligence
Automatic Identification of Perceptually Important Regions in an Image
ICPR '98 Proceedings of the 14th International Conference on Pattern Recognition-Volume 1 - Volume 1
A Coherent Computational Approach to Model Bottom-Up Visual Attention
IEEE Transactions on Pattern Analysis and Machine Intelligence
Axiomatic approach to computational attention
Pattern Recognition
| Hi-index | 0.00 |
Is there any advertisement in a particular dataset more visually efficient than the rest? Here we propose that advertisement images may be rank ordered based on their important information visibility using computational attention. For each one of the advertisement images we first compute a multi-bitrate attention map following a rational model of computational attention. Next, based on the attention map, we calculate the average attention score, for each bitrate, within the areas of interest either provided by the publicist or by the use of automated detection. A high value of the mean attention within the areas of interest at any reconstruction fidelity corresponds to a high saliency of these areas. Thus, for each advertisement, we calculate a rate-attention curve as given by the normalized mean attention score within the areas of interest across bitrates. Each image is decoded at different bitrates of picture quality using a coding method. Unsupervised learning can then be used to perform the clustering of the advertisements into subsets so that images in the same cluster are similar in the rate-attention sense. In the experiments one advertisement has appeared to be more visually efficient than the rest of images in a dataset of example.