Finding Distractors In Images
Computer Vision and Pattern Recognition (CVPR), June 2015
Distractor Removal. From left to right: (1) Original image. (2) Normalized average ground-truth annotation. (3) Order of segments as predicted by our algorithm. (4) Distractor removal result. Segment selected for removal are shown in red. Notice how these correlate with the ground-truth annotation.
Abstract
We propose a new computer vision task we call "distractor prediction." Distractors are the regions of an image that draw attention away from the main subjects and reduce the overall image quality. Removing distractors -- for example, using in-painting -- can improve the composition of an image. In this work we created two datasets of images with user annotations to identify the characteristics of distractors. We use these datasets to train an algorithm to predict distractor maps. Finally, we use our predictor to automatically enhance images.
Download
- Paper (preprint) (4MB PDF)
- Video (36.5MB MP4)
- Data (91.6MB ZIP)
- Code
- More
Citation
Ohad Fried, Eli Shechtman, Dan B Goldman, and Adam Finkelstein.
"Finding Distractors In Images."
Computer Vision and Pattern Recognition (CVPR), June 2015.
BibTeX
@inproceedings{Fried:2015:FDI,
author = "Ohad Fried and Eli Shechtman and Dan B Goldman and Adam Finkelstein",
title = "Finding Distractors In Images",
booktitle = "Computer Vision and Pattern Recognition (CVPR)",
year = "2015",
month = jun
}