Optimized cutting plane algorithm for support vector machines

Franc, Vojtĕch; Sonnenburg, Sören

doi:10.1145/1390156.1390197

Cited by 109 publications

(119 citation statements)

References 12 publications

Supporting

Mentioning

118

Contrasting

Order By: Relevance

“…b) Support Vector Machine (SVM) using kernel function with linear mapping [33], which is composed with the m 1 (x) and a randomly chosen subset from the trained data. It is trained using a fast cutting plane algorithm [35].…”

Section: Voxel Classificationmentioning

confidence: 99%

Line filtering for surgical tool localization in 3D ultrasound images

Uhercik¹,

Kybic²,

Zhao³

et al. 2013

Computers in Biology and Medicine

View full text Add to dashboard Cite

We present a method for automatic surgical tool localization in 3D ultrasound images based on line filtering, voxel classification and model fitting. A possible application is to provide assistance for biopsy needle or micro-electrode insertion, or a robotic system performing this insertion. The line filtering method is first used to enhance the contrast of the 3D ultrasound image, then a classifier is chosen to separate the tool voxels, in order to reduce the number of outliers. The last step is a RANSAC model fitting. Experimental results on several different datasets demonstrate that the failure rate of the proposed method is lower than for preceding methods, with similar localization accuracy, at the expense of a modest increase in computational effort. Even though some of the line filtering methods are a little slower than other methods, its robustness and accuracy are strongly proposed and it is worth to develop a system for clinique applications.

show abstract

Section: Voxel Classificationmentioning

confidence: 99%

Line filtering for surgical tool localization in 3D ultrasound images

Uhercik¹,

Kybic²,

Zhao³

et al. 2013

Computers in Biology and Medicine

View full text Add to dashboard Cite

show abstract

“…Such landmark positions are used for normalized face image construction, which is used in the next recognition step. The problem of gender and identity estimation are classification problems, for which we use multi-class Support Vector Machine (SVM) classifier, more precisely the implementation done by Franc [10]. The classifier uses image features, computed from normalized face images, based on Local Binary Patterns (LBP).…”

Section: Face Recognitionmentioning

confidence: 99%

Online Speaker Adaptation of an Acoustic Model Using Face Recognition

Campr

Pražák²,

Psutka³

2013

Text, Speech, and Dialogue

View full text Add to dashboard Cite

Abstract. We have proposed and evaluated a novel approach for online speaker adaptation of an acoustic model based on face recognition. Instead of traditionally used audio-based speaker identification we investigated the video modality for the task of speaker detection. A simulated on-line transcription created by a Large-Vocabulary Continuous Speech Recognition (LVCSR) system for online subtitling is evaluated utilizing speaker independent acoustic models, gender dependent models and models of particular speakers. In the experiment, the speaker dependent acoustic models were trained offline, and are switched online based on the decision of a face recognizer, which reduced Word Error Rate (WER) by 12% relatively compared to speaker independent baseline system.

show abstract

“…The modified algorithm is called the optimized cutting plane algorithm (OCAS) [48], and it was shown that the number of iterations needed for convergence is the same order as that in [47]. On the other hand, OCAS is reported to be much faster in numerical experiments than the algorithm proposed in [47].…”

Section: ) the Empirical Risk Is Expressed Asmentioning

confidence: 99%

Recent Advances and Trends in Large-Scale Kernel Methods

Kashima

Idé

Kato

et al. 2009

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

SUMMARYKernel methods such as the support vector machine are one of the most successful algorithms in modern machine learning. Their advantage is that linear algorithms are extended to non-linear scenarios in a straightforward way by the use of the kernel trick. However, naive use of kernel methods is computationally expensive since the computational complexity typically scales cubically with respect to the number of training samples. In this article, we review recent advances in the kernel methods, with emphasis on scalability for massive problems.

show abstract

Optimized cutting plane algorithm for support vector machines

Cited by 109 publications

References 12 publications

Line filtering for surgical tool localization in 3D ultrasound images

Line filtering for surgical tool localization in 3D ultrasound images

Online Speaker Adaptation of an Acoustic Model Using Face Recognition

Recent Advances and Trends in Large-Scale Kernel Methods

Contact Info

Product

Resources

About