Image Tag Completion by Noisy Matrix Recovery

Feng, Zheyun; Feng, Songhe; Jin, Rong; Jain, Anil K.

doi:10.1007/978-3-319-10584-0_28

Cited by 43 publications

(51 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The key difference is that the main objective of our model is to segment a set of test images without any labels rather than segmenting the weakly-labelled training images. Moreover, our work is also related to image tagging works such as [20], [21], [43] which learn a model from noisy user-provided image-level labels as well. However, the problem tackled here is harder as we aim to simultaneously segment images and label each segment, rather than labelling the whole images.…”

Section: Related Workmentioning

confidence: 99%

“…This problem is further compounded when the image-level weak labels also contain noise -particularly when the user-provided tags (e.g. those from Flickr) are used as labels for model training [20], [21]. In addition, many existing methods [11], [12], [17]- [19] need to use predicted image-level labels for the test images as model input, which again are noisy due to imperfect prediction.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Learning from Weak and Noisy Labels for Semantic Segmentation

Xiang

et al. 2017

IEEE Trans. Pattern Anal. Mach. Intell.

113

View full text Add to dashboard Cite

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Learning from Weak and Noisy Labels for Semantic Segmentation

Xiang

et al. 2017

IEEE Trans. Pattern Anal. Mach. Intell.

113

View full text Add to dashboard Cite

“…Meanwhile, users often avoid assigning tags to images, making image tags incomplete. To deal with noisy and incomplete tags, several completion methods [2,3,7] have been recently proposed, Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than ACM must be honored.…”

Section: Introductionmentioning

confidence: 99%

“…Tags are provided by users in the form of free text and they are usually imprecise, containing noise to efficiently describe the visual content of images [2,7]. Meanwhile, users often avoid assigning tags to images, making image tags incomplete.…”

Section: Introductionmentioning

confidence: 99%

Probabilistic Matrix Factorization With Semantic And Visual Neighborhoods For Image Tag Completion

Rafailidis

2015

Proceedings of the 5th ACM on International Conference on Multimedia Retrieval

View full text Add to dashboard Cite

We present an image tag completion method, namely PMF-SVN, where the key idea is to exploit images' Semantically and Visually similar Neighborhoods (SVNs) in the learning process of a Probabilistic Matrix Factorization (PMF) framework. We propose a two-step SVN formation algorithm that can generate an image set with the images being both visually and semantically similar. Furthermore, we introduce an efficient way to incorporate the formed SVNs into the learning process of PMF, under the constraint that the latent features of each image are averaged by the features of the images that belong to its SVN. In our experiments with benchmark datasets, we show that the proposed PMF-SVN method outperforms competitive baselines, in terms of completion accuracy, by efficiently capturing the semantical and visual associations between images and tags in SVNs.

show abstract

“…Image tag assignment strives to assign a number of tags related to the image content to unlabeled images [2,20,21,22]. Image tag refinement aims to remove irrelevant tags from the initial tag list and enrich it with novel, yet relevant tags [23,24,25]. Inspired by the technologies used in image tagging, we will use the co-regularization framework, as shown in figure 1, to fast and automatically tag large natural sounds.…”

Section: Introductionmentioning

confidence: 99%

Fast tagging of natural sounds using marginal co-regularization

Huang

Jackson

et al. 2017

2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

Automatic and fast tagging of natural sounds in audio collections is a very challenging task due to wide acoustic variations, the large number of possible tags, the incomplete and ambiguous tags provided by different labellers. To handle these problems, we use a co-regularization approach to learn a pair of classifiers on sound and text. The first classifier maps low-level audio features to a true tag list. The second classifier maps actively corrupted tags to the true tags, reducing incorrect mappings caused by low-level acoustic variations in the first classifier, and to augment the tags with additional relevant tags. Training the classifiers is implemented using marginal co-regularization, pair of which draws the two classifiers into agreement by a joint optimization. We evaluate this approach on two sound datasets, Freefield1010 and Task4 of DCASE2016. The results obtained show that marginal co-regularization outperforms the baseline GMM in both efficiency and effectiveness.

show abstract

Image Tag Completion by Noisy Matrix Recovery

Cited by 43 publications

References 24 publications

Learning from Weak and Noisy Labels for Semantic Segmentation

Learning from Weak and Noisy Labels for Semantic Segmentation

Probabilistic Matrix Factorization With Semantic And Visual Neighborhoods For Image Tag Completion

Fast tagging of natural sounds using marginal co-regularization

Contact Info

Product

Resources

About