Tag Completion for Image Retrieval

Wu, Lei; Jin, Rong; Jain, Anil K.

doi:10.1109/tpami.2012.124

Cited by 211 publications

(4 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although both positive unlabelled multi-label learning (Sun et al, 2010;Bucak et al, 2011;Wei et al, 2018;Wu et al, 2013;Kanehira and Harada, 2016;Teisseyre, 2021) as well as the problem of feature selection in multi-label classification (Pereira et al, 2018;Kashef et al, 2018;Lee and Kim, 2017) have attracted close attention, a combination of these two problems, to the best of our knowledge, remains an unexplored area. In this paper, we focus on penalized empirical risk minimization frameworks.…”

Section: 2mentioning

confidence: 99%

Joint feature selection and classification for positive unlabelled multi-label data using weighted penalized empirical risk minimization

Teisseyre

2022

View full text Add to dashboard Cite

We consider the positive-unlabelled multi-label scenario in which multiple target variables are not observed directly. Instead, we observe surrogate variables indicating whether or not the target variables are labelled. The presence of a label means that the corresponding variable is positive. The absence of the label means that the variable can be either positive or negative. We analyze embedded feature selection methods based on two weighted penalized empirical risk minimization frameworks. In the first approach, we introduce weights of observations. The idea is to assign larger weights to observations for which there is a consistency between the values of the true target variable and the corresponding surrogate variable. In the second approach, we consider a weighted empirical risk function which corresponds to the risk function for the true unobserved target variables. The weights in both the methods depend on the unknown propensity score functions, whose estimation is a challenging problem. We propose to use very simple bounds for the propensity score, which leads to relatively simple forms of weights. In the experiments we analyze the predictive power of the methods considered for different labelling schemes.

show abstract

Section: 2mentioning

confidence: 99%

Joint feature selection and classification for positive unlabelled multi-label data using weighted penalized empirical risk minimization

Teisseyre

2022

View full text Add to dashboard Cite

show abstract

“…Tag Refinement: The problem of tag refinement aims at removing noisy tags from images while adding more relevant ones [22]. This line of work appears in the literature also as tag completion [11,46] or image re-tagging [24,25]. Most of this work uses only nouns as tags and disregards word ambiguity.…”

Section: Related Workmentioning

confidence: 99%

“…This assumption is often invalid. This body of work employs a variety of methods, including metric learning [14], matrix completion [11,46,50], latent topic models [47], and more.…”

Section: Related Workmentioning

confidence: 99%

Visir

Chowdhury

Tandon

Ferhatosmanoğlu

et al. 2018

Proceedings of the Eleventh ACM International Conference on Web Search and Data Mining

View full text Add to dashboard Cite

The social media explosion has populated the Internet with a wealth of images. There are two existing paradigms for image retrieval: 1) content-based image retrieval (CBIR), which has traditionally used visual features for similarity search (e.g., SIFT features), and 2) tag-based image retrieval (TBIR), which has relied on user tagging (e.g., Flickr tags). CBIR now gains semantic expressiveness by advances in deep-learning-based detection of visual labels. TBIR benefits from query-and-click logs to automatically infer more informative labels. However, learning-based tagging still yields noisy labels and is restricted to concrete objects, missing out on generalizations and abstractions. Click-based tagging is limited to terms that appear in the textual context of an image or in queries that lead to a click. This paper addresses the above limitations by semantically refining and expanding the labels suggested by learning-based object detection. We consider the semantic coherence between the labels for different objects, leverage lexical and commonsense knowledge, and cast the label assignment into a constrained optimization problem solved by an integer linear program. Experiments show that our method, called VISIR, improves the quality of the state-of-the-art visual labeling tools like LSDA and YOLO.

show abstract

“…However, many user-provided tags are incomplete or inaccurate in describing the visual content of images [1], making them difficult to be utilized for tasks such as tag based image retrieval and tag recommendation. Therefore, automatic image annotation is one of the challenge problems.…”

Section: Introductionsmentioning

confidence: 99%

Manifold Kernel Metric Learning for Larger-Scale Image Annotation

Guo

2015

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

SUMMARYAn appropriate similarity measure between images is one of the key techniques in search-based image annotation models. In order to capture the nonlinear relationships between visual features and image semantics, many kernel distance metric learning(KML) algorithms have been developed. However, when challenged with large-scale image annotation, their metrics can't explicitly represent the similarity between image semantics, and their algorithms suffer from high computation cost. Therefore, they always lose their efficiency. In this paper, we propose a manifold kernel metric learning (M KML) algorithm. Our M KML algorithm will simultaneously learn the manifold structure and the image annotation metrics. The main merit of our M KML algorithm is that the distance metrics are builded on image feature's interior manifold structure, and the dimensionality reduction on manifold structure can handle the high dimensionality challenge faced by KML. Final experiments verify our method's efficiency and effectiveness by comparing it with state-of-the-art image annotation approaches. key words: image annotation, kernel distance metric learning, manifold structure learning

show abstract

Tag Completion for Image Retrieval

Cited by 211 publications

References 48 publications

Joint feature selection and classification for positive unlabelled multi-label data using weighted penalized empirical risk minimization

Joint feature selection and classification for positive unlabelled multi-label data using weighted penalized empirical risk minimization

Visir

Manifold Kernel Metric Learning for Larger-Scale Image Annotation

Contact Info

Product

Resources

About