Training Deep Networks to be Spatially Sensitive
Open Access
- 1 October 2017
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 5669-5678
Abstract
In many computer vision tasks, for example saliency prediction or semantic segmentation, the desired output is a foreground map that predicts pixels where some criteria is satisfied. Despite the inherently spatial nature of this task commonly used learning objectives do not incorporate the spatial relationships between misclassified pixels and the underlying ground truth. The Weighted F-measure, a recently proposed evaluation metric, does reweight errors spatially, and has been shown to closely correlate with human evaluation of quality, and stably rank predictions with respect to noisy ground truths (such as a sloppy human annotator might generate). However it suffers from computational complexity which makes it intractable as an optimization objective for gradient descent, which must be evaluated thousands or millions of times while learning a model's parameters. We propose a differentiable and efficient approximation of this metric. By incorporating spatial information into the objective we can use a simpler model than competing methods without sacrificing accuracy, resulting in faster inference speeds and alleviating the need for pre/post-processing. We match (or improve) performance on several tasks compared to prior state of the art by traditional metrics, and in many cases significantly improve performance by the weighted F-measure.Keywords
All Related Versions
This publication has 23 references indexed in Scilit:
- Deep Contrast Learning for Salient Object Detection2016
- Deep networks for saliency detection via local estimation and global search2015
- Saliency detection by multi-context deep learning2015
- Saliency Optimization from Robust Background Detection2014
- Category-Independent Object-Level Saliency Detection2013
- Hierarchical Saliency Detection2013
- Salient Object Detection: A Discriminative Regional Feature Integration Approach2013
- Learning to predict where humans look2009
- Rapid object detection using a boosted cascade of simple features2005
- Maximum expected F-measure training of logistic regression models2005