Large-Scale Image Annotation by Efficient and Robust Kernel Metric Learning

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

doi:10.1109/iccv.2013.203

Cited by 46 publications

(25 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Equation (3) degenerates a standard RKML problem, and can be solved use the method [22]. Then fixed T , and solve U.…”

Section: The Manifold Kernel Matric Learning (M Kml) Algorithmmentioning

confidence: 99%

“…, n, the regularization parameter α and β, the heat kernel L(L = diag(W T I)−W), properly initialize the matrices U 0 and T 0 , then, set t ← 0 Step1: Calculate the chi-squared kernel K of the training samples, and t = t + 1 ; Step2: Fixed U, and solve T . Equation (9) can be solved by the method [22], and get the optimization T t ; Step3: Fixed T , and solve U. Equation (10) can be solved to obtain the optimal solution by LARS, and get the optimization U t ; Step4: If U t − U t−1 2 F and T t − T t−1 2 F does not decrease, or the iteration number t is larger than a predefined threshold, then exit.…”

Section: The Manifold Kernel Matric Learning (M Kml) Algorithmmentioning

confidence: 99%

“…However, images always have a nonlinear relationship and multi-modal patterns [16]- [19]. Kernel Metric Learning (KML) [16], [18], [22] has been proposed to overcome this problem. The key idea is to map the data from the original vector space into the kernel nonlinear spaces.…”

Section: Introductionsmentioning

confidence: 99%

“…Finally, the high dimensionality of KML may lead to the over-fitting of training data [21]. Therefore, Feng [22] proposed a robust kernel metric learning(RKML) algorithm based on the regression technique that is able to directly utilize image annotations. RKML model the image annotation as a following optimization problem:…”

Section: Introductionsmentioning

confidence: 99%

See 3 more Smart Citations

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

“…Equation (3) degenerates a standard RKML problem, and can be solved use the method [22]. Then fixed T , and solve U.…”

Section: The Manifold Kernel Matric Learning (M Kml) Algorithmmentioning

confidence: 99%

Section: The Manifold Kernel Matric Learning (M Kml) Algorithmmentioning

confidence: 99%

Section: Introductionsmentioning

confidence: 99%

Section: Introductionsmentioning

confidence: 99%

See 2 more Smart Citations

Manifold Kernel Metric Learning for Larger-Scale Image Annotation

Guo

2015

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

show abstract

“…Numerous metric learning methods have been proposed for a variety of computer vision applications such as face verification [10,17,27], object classification [26], image annotation [7,33], and visual tracking [21], etc. The information conveyed by training data can be generally represented as triplet and pairwise constraints.…”

Section: Introductionmentioning

confidence: 99%

Shrinkage Expansion Adaptive Metric Learning

Wang

Zuo

Zhang

et al. 2014

Computer Vision – ECCV 2014

View full text Add to dashboard Cite

Abstract. Conventional pairwise constrained metric learning methods usually restrict the distance between samples of a similar pair to be lower than a fixed upper bound, and the distance between samples of a dissimilar pair higher than a fixed lower bound. Such fixed bound based constraints, however, may not work well when the intra-and inter-class variations are complex. In this paper, we propose a shrinkage expansion adaptive metric learning (SEAML) method by defining a novel shrinkage-expansion rule for adaptive pairwise constraints. SEAML is very effective in learning metrics from data with complex distributions. Meanwhile, it also suggests a new rule to assess the similarity between a pair of samples based on whether their distance is shrunk or expanded after metric learning. Our extensive experimental results demonstrated that SEAML achieves better performance than state-of-the-art metric learning methods. In addition, the proposed shrinkage-expansion adaptive pairwise constraints can be readily applied to many other pairwise constrained metric learning algorithms, and boost significantly their performance in applications such as face verification on LFW and PubFig databases.

show abstract