M.S. Thesis

M.S. Thesis

Evrim Fer, Testing Natural Selection on Polygenic Trait-Associated Alleles in Anatolia Using Neolithic and Present-Day Human Genomes

Neolithic transition which has started approximately 10,000 year ago in west Eurasia and created a major shifts in human life. In many studies, strong selection signals on the genes related to changes in life-style were determined. With the advent of archeogenomics studies, those adaptations have also been supported using ancient DNA. Here, polygenic adaptations in Anatolia were investigated by comparing Neolithic (n=36) and modern-day (n=16) genome sequences for 40 polygenic traits potentially associated with diet, immunity and other related complex traits using pairwise FST and population branch statistics (PBS) method which provides a genome-wide selection analysis including a distant population.

Date: 29.07.2019 Place: A-212

English

Orhun Olgun, Entropy-based Direction-of-Arrival Estimation Methods for Rigid Spherical Microphone Arrays

The work reported in thesis investigates direction-of-arrival estimation methods for sound sources and proposes novel method called HiGRID and its extensions with DPD test and EB-MUSIC. Proposed methods are based on spherical harmonic decomposition and rigid spherical microphone arrays enables trivial calculation of spherical harmonic decomposition of sound fields. Evaluation of proposed methods are presented in terms of direction-of-arrival estimation errors.

Date: 26.07.2019 PlaceA-212

Orhun Olgun, Entropy-based Direction-of-Arrival Estimation Methods for Rigid Spherical Microphone Arrays

English

Timuçin Berk Atalay, Scattering Delay Networks with Aperture Size Control for Simulating Coupled Volume Acoustics

Artificial reverberators provide a computationally viable alternative to full-scale room acoustics simulation methods for games and extended reality applications. Scattering delay network (SDN) is a relatively recent artificial reverberator that allows direct parametric control over the geometry of a simulated rectangular enclosure. We propose a new model called coupled volume scattering delay networks (CV-SDN) which extends the classic SDN structure for enclosures coupled via an aperture. The extension allows independent control of acoustical properties of volumes and the size of the connecting aperture. The utility of the proposed method is demonstrated by the energy decay curves obtained from impulse responses.

Date: 26.07.2019 PlaceA-212


English

Ozan Emirhan Bayyurt, Designing and Implementing a Game Development Framework for Interactive Stories and Role Playing Games

Video games are a great medium for storytelling. This thesis proposes a game development framework to help developers create role-playing games in a much easier and quicker fashion. With its modular structure, the proposed framework offers developers a highly scalable game development environment.

Date: 23.07.2019 10:00 Place: Computer Engineering Building A-105

Ozan Emirhan Bayyurt, İnteraktif Hikaye Ve Rol Yapma Oyunlarına Odaklı Bir Oyun Geliştirme Sisteminin Tasarlanması Ve İmplementasyonu

English

Necati Çağatay Gürsoy, Cross-age Effect In Artificial Neural Networks: A Study On Facial Age Recognition Bias In Artificial Neural Networks

Simply put cross-age effect or own-age bias is the phenomenon that when guessing the age of individuals from their faces, it is claimed that the age of the individual who is guessing the facial age effects the guessing process and the effect is towards to the age of the one who is guessing. In our study; the phenomenon is investigated in humans to observe if such effect exists, then it is tested if the error could be reduced via utilizing convolutional neural networks on face images and classify the face images with respect to their ages.

Date: 27.06.2019 16:30 Place: A-212

English

Salih Fırat Canpolat, A Novel Approach To Emotion Recognition In Voice: A Convolutional Neural Network Approach And Grad-Cam Generation

The study deals with the emotion recognition problem in Turkish single word pronunciations through the image recognition perspective. The CNN approach allowed us to use spectrograms as the training material. The resulting model has a feasible predictive power and shares trends with the human judges: it is robust to the changes in the sound signal at higher frequencies and generally performed better than the judges except for the 500-8000 hertz band where the human judges did not lose any significant predictive power. The model proved to be a feasible explanation of human assessment of emotions by failing at 500-8000 hertz band where the humans remained robust.

Date: 27.06.2019 13:30   Place: A-108

English

Pages

Subscribe to RSS - M.S. Thesis