Pixel Club: On Saliency Detection & its Evaluation

Speaker:
Ran Margolin (EE, Technion)
Date:
Tuesday, 15.7.2014, 11:30
Place:
EE Meyer Building 1061

I will begin by discussing what makes a patch salient. Most previous work assert that distinctness is the dominating factor. The difference between the various algorithms is in the way they compute distinctness. Some focus on the patterns, others on the colors, and several add high-level cues and priors.

We propose a simple, yet powerful, algorithm that integrates these three factors. Our key contribution is a novel and fast approach to compute pattern distinctness. We rely on the inner statistics of the patches in the image for identifying unique patterns.

We show that our approach outperforms all state-of-the-art methods on the five most commonly-used datasets. Then, I will revisit the common measures of evaluation of saliency detection and foreground maps. We show that the most commonly-used measures for evaluating both saliency maps and foreground maps do not always provide a reliable evaluation. Several measures have been suggested to evaluate the accuracy of these maps. This includes the Area-Under-the-Curve measure, the Average-Precision measure, the F-measure, and the evaluation measure of the PASCAL VOC segmentation challenge. We start by identifying three causes of inaccurate evaluation.

We then propose a new measure that amends these flaws. An appealing property of our measure is being an intuitive generalization of the F-measure. Finally we propose four meta-measures to compare the adequacy of evaluation measures. We show via experiments that our novel measure is preferable.

Back to the index of events