Bilge Gunsel

Learn More
This paper proposes a single-channel audio source decomposition method that integrates perceptual quality criteria into source separation. Unlike the existing methods, the proposed method applies a perceptually weighted non-negative matrix factorization on log-frequency spectrogram of the mixed signal. The weights are adaptively calculated for each critical(More)
The sea target detection problem from thermal (IR) images is solved by using statistical classification methods. Background modelling is achieved via principle component analysis (PCA) followed by a two-class Bayes classification step, i.e., target or sea. A wavelet-denoising block is added to the system resulting in a significant increase in the detection(More)
This paper proposes an adaptive audio watermarking algorithm that allows watermark embedding at low watermark-to-signal ratio levels. It is realized by modifying the MPEG Layer 1 Model 1 psychoacoustic masking technique in an iterative way, while preserving inaudibility. The developed watermarking algorithm does not require the original audio file at the(More)
This paper presents a system for emotion recognition from audio. The emotion recognition problem is described as a four class classification problem in a 17 dimensional feture space. Currently, considered emotions are: Anger, happiness, sadness and unemotional. A SVM classifier with Gaussian RBF kernel is used for the classification. Experimental results(More)
This paper compares two audio spotting methods: a feature-based audio classification method and a hashing method for audio fingerprinting. Moreover, the music genre determination performance of the methods is investigated. In this context, advantages and disadvantages of these methods are discussed and the audio identification performance of both methods(More)
As the amount of traffic generated by next generation mobile services keeps increasing, the operators face the scalability challenges of VoIP protocols. This paper presents an analysis tool to evaluate the effect of SIP retransmission mechanism on the SIP server overloading. Our tool is capable of creating various arrival and departure processes using(More)
Recently introduced variable rate particle filters (VRPF), utilizing semi-Markov models for maneuvering target tracking obtained superior performance compared to the standard Markov models [1]. However, degeneracy problem commonly encountered in particle filtering also arises in VRPF algorithm. In this work, regularization technique used in standard(More)
This paper introduces ArtHistorian, an art painting indexing system designed for automatic generation of dynamic painting presentations. The developed system aims personalized content browsing with a classification based indexing and query method implemented from the art historians’ perspective. ArtHistorian represents the visual content of paintings by a(More)
Our aim in this paper is to illustrate the effectiveness of the Dirichlet Process Mixture (DPM) model for emotional speech class density estimation when the number of Gauss mixture components are unknown. The problem is modeled as a two-class classification problem where the classes are anger and-no-anger. Performance of the algorithm is evaluated on the(More)