Semi-supervised dimension reduction approaches integrating global and local pattern information

Date

2019-02-01

Author

Sakarya, Ufuk

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

205
views

0
downloads

Dimension reduction is an important research area in pattern recognition. Use of both supervised and unsupervised data can be an advantage in the case of lack of labeled training data. Moreover, use of both global and local pattern information can contribute classification performances. Therefore, four important primary components are essential to design a well-performed semi-supervised dimension reduction approach: global pattern modeling by a supervised manner, local pattern modeling by a supervised manner, global pattern modeling by an unsupervised manner, and local pattern modeling by an unsupervised manner. These primary components are integrated into two proposed methods. The first is the semi-supervised global-local linear discriminant analysis, and the second is the semi-supervised global-local maximum margin criterion. The proposed methods are examined in object recognition and hyperspectral image classification. According to the experimental results, the promising results are obtained against to comparative semi-supervised methods.

Subject Keywords

Signal Processing, Electrical and Electronic Engineering

URI

https://hdl.handle.net/11511/63659

Journal

SIGNAL IMAGE AND VIDEO PROCESSING

DOI

https://doi.org/10.1007/s11760-018-1342-5

Collections

Unclassified, Article

Suggestions

OpenMETU
Core

Convolutional neural networks analysed via inverse problem theory and sparse representations Tarhan, Cem; Akar, Gözde (Institution of Engineering and Technology (IET), 2019-04-01) Inverse problems in imaging such as denoising, deblurring, superresolution have been addressed for many decades. In recent years, convolutional neural networks (CNNs) have been widely used for many inverse problem areas. Although their indisputable success, CNNs are not mathematically validated as to how and what they learn. In this study, the authors prove that during training, CNN elements solve for inverse problems which are optimum solutions stored as CNN neuron filters. They discuss the necessity of mu...
Extended Target Tracking Using Polynomials With Applications to Road-Map Estimation Lundquist, Christian; Orguner, Umut; Gustafsson, Fredrik (Institute of Electrical and Electronics Engineers (IEEE), 2011-01-01) This paper presents an extended target tracking framework which uses polynomials in order to model extended objects in the scene of interest from imagery sensor data. State-space models are proposed for the extended objects which enables the use of Kalman filters in tracking. Different methodologies of designing measurement equations are investigated. A general target tracking algorithm that utilizes a specific data association method for the extended targets is presented. The overall algorithm must always ...
Multiple target localization & data association for frequency-only widely separated MIMO radar Kalkan, Yilmaz; Baykal, Buyurman (Elsevier BV, 2014-02-01) The multiple target localization using only Doppler frequencies in Multiple-Input Multiple-Output (MIMO) radar is a problem of great interest. The Doppler frequencies include all relevant information about the location, velocity and direction of the targets. Hence, these frequencies can be used efficiently for data association and target localization. The MIMO configuration and the frequency diversity of the system enable us to easily determine the number of moving targets using only the Doppler frequencies...
Maximum likelihood estimation of transition probabilities of jump Markov linear systems Orguner, Umut (Institute of Electrical and Electronics Engineers (IEEE), 2008-10-01) This paper describes an online maximum likelihood estimator for the transition probabilities associated with a jump Markov linear system (JMLS). The maximum likelihood estimator is derived using the reference probability method, which exploits an hypothetical probability measure to find recursions for complex expectations. Expectation maximization (EM) procedure is utilized for maximizing the likelihood function. In order to avoid the exponential increase in the number of statistics of the optimal EM algori...
Block-based methods for the reconstruction of finite-length signals from nonuniform samples Tuncer, Temel Engin (Institute of Electrical and Electronics Engineers (IEEE), 2007-02-01) Two novel block-based algorithms are presented for the reconstruction of uniform samples given the nonuniform samples. The first algorithm uses a sinc interpolator whereas the second one uses a DFT-based interpolator. It is shown that the proposed algorithms are stable and the error due to noise and sampling jitter is bounded by the corresponding error norms of noise and jitter, respectively. We show that both of the block-based algorithms provide nearly perfect reconstruction for a class of practically tim...

Citation Formats

U. Sakarya, “Semi-supervised dimension reduction approaches integrating global and local pattern information,” SIGNAL IMAGE AND VIDEO PROCESSING, pp. 171–178, 2019, Accessed: 00, 2020. [Online]. Available: https://hdl.handle.net/11511/63659.