Avci, Umut
Loading...
Name Variants
Job Title
Dr.Öğr.Üyesi
Email Address
Main Affiliation
01.01.09.07. Yazılım Mühendisliği Bölümü
Status
Current Staff
Website
ORCID ID
Scopus Author ID
Turkish CoHE Profile ID
Google Scholar ID
WoS Researcher ID
Files
Sustainable Development Goals
SDG data is not available
Documents
14
Citations
93
h-index
5
Documents
9
Citations
56
Scholarly Output
9
Articles
3
Views / Downloads
0/1
Supervised MSc Theses
0
Supervised PhD Theses
0
WoS Citation Count
5
Scopus Citation Count
20
Patents
0
Projects
0
WoS Citations per Publication
0.56
Scopus Citations per Publication
2.22
Open Access Source
2
Supervised Theses
0
| Journal | Count |
|---|---|
| 14th IEEE International Conference on Automatic Face and Gesture Recognition FG 2019 | 1 |
| 21st International Conference on Speech and Computer SPECOM 2019 | 1 |
| 22nd International Conference on Speech and Computer SPECOM 2020 | 1 |
| 4th International Conference on Intelligent and Fuzzy Systems (INFUS) | 1 |
| 5th International Congress on Human-Computer Interaction Optimization and Robotic Applications HORA 2023 | 1 |
Current Page: 1 / 2
Scopus Quartile Distribution
Quartile distribution chart data is not available
Competency Cloud
9 results
Scholarly Output Search Results
Now showing 1 - 9 of 9
Conference Object A Comparative Study of Artificial Intelligence Based Methods for Abnormal Pattern Identification in SPC(SPRINGER INTERNATIONAL PUBLISHING AG, 2022) Umut Avci; Onder Bulut; Ayhan Ozgur Toy; Toy, Ayhan Ozgur; Bulut, Onder; Avci, Umut; C Kahraman; AC Tolga; SC Onar; S Cebi; B Oztaysi; IU SariStatistical process control techniques have been used to detect any assignable cause that may result in a lower quality. Among these techniques is the identification of any abnormal patterns that may indicate the presence of an assignable cause. These abnormal patterns may be in the form of steady movement in one direction i.e. trends, an instantaneous change in the process mean i.e. sudden shift, a series of high observations followed by a series of low observations i.e. cycles. As long as we can classify the observed data the decision maker can decide on actions to be performed to ensure quality standards and planning for interventions. In identification of these abnormal patterns rather than relying on human element intelligent tools have been proposed in the literature. We attempt to provide a comparative study of various classification algorithms used for pattern identification in statistical process control. We specifically consider six different types of patterns to classify. These different types are: (1) Normal (2) Upward trend (3) Downward trend (4) Upward shift (5) Downward shift (6) Cyclic. A recent trend in classification is to use deep neural networks (DNNs). However due to the design complexity of DNNs alternative classification methods should also be considered. Our focus on this study is to compare traditional classification methods to a recent DNN solution in the literature in terms of their efficiencies. Our numerical study indicates that basic classification algorithms perform relatively well in addition to their structural advantages.Article Citation - WoS: 1Citation - Scopus: 1A Pattern Mining Approach for Improving Speech Emotion Recognition(WORLD SCIENTIFIC PUBL CO PTE LTD, 2022) Umut Avci; Avci, UmutSpeech-driven user interfaces are becoming more common in our lives. To interact with such systems naturally and effectively machines need to recognize the emotional states of users and respond to them accordingly. At the heart of the emotion recognition research done to this end lies the emotion representation that enables machines to learn and predict emotions. Speech emotion recognition studies use a wide range of low-to-high-level acoustic features for representation purposes such as LLDs their functionals and BoAW. In this paper we present a new method for extracting a novel set of high-level features for classifying emotions. For this purpose we (1) reduce the dimension of discrete-time speech signals (2) perform a quantization operation on the new signals and assign a distinct symbol to each quantization level (3) use the symbol sequences representing the signals to extract discriminative patterns that are capable of distinguishing different emotions from each other and (4) generate a separate set of features for each emotion from the extracted patterns. Experimental results show that pattern features outperform Energy Voicing MFCC Spectral and RASTA feature sets. We also demonstrate that combining the pattern-based features and the acoustic features further improves the classification performance.Conference Object Citation - WoS: 1Citation - Scopus: 2A pattern mining approach in feature extraction for emotion recognition from speech(Springer Verlag service@springer.de, 2019) Umut Avci; Gamze Akkurt; Devrim Ünay; Unay, Devrim; Avci, Umut; Akkurt, Gamze; A.A. Salah , A.A. Salah , A. Karpov , R. PotapovaWe address the problem of recognizing emotions from speech using features derived from emotional patterns. Because much work in the field focuses on using low-level acoustic features we explicitly study whether high-level features are useful for classifying emotions. For this purpose we convert a continuous speech signal to a discretized signal and extract discriminative patterns that are capable of distinguishing distinct emotions from each other. Extracted patterns are then used to create a feature set to be fed into a classifier. Experimental results show that patterns alone are good predictors of emotions. When used to build a classifier pattern features achieve accuracy gains up to 25% compared to state-of-the-art acoustic features. © 2019 Elsevier B.V. All rights reserved.Conference Object Citation - WoS: 1Citation - Scopus: 2Analyzing group performance in small group interaction: Linking personality traits and group performance through the verbal content(Institute of Electrical and Electronics Engineers Inc., 2019) Umut Avci; Oya Aran; Aran, Oya; Avci, UmutIn this paper we investigate the link between the personality traits and group performance in terms of the verbal content. We further study the variability in the verbal interaction between different performance groups. Towards this goal we extract topics representing the content of meetings as well as term-frequencies of items that play a critical role in the decision task. We use a dataset where each group performs the winter survival task in which the task is to decide on the ranking of different items with respect to the importance of each item for their survival. In the experiments we contrast the ranking of items with respect to their term frequencies and compare the differences between topics both for distinct personality traits and group performances. Results of the term-frequency based approach show that influential people put correct emphasis on items more than dominant people. The topic-based method reveals that influential people consider the majority of items by providing usage instructions for alternative scenarios and that dominant people focus only on a small subset of items by stressing their significance. High-performance groups assess items in a similar manner to influential and dominant people i.e. a wide range of items are considered and their importance is explained. Low-performance groups on the other hand concentrate on the situation they are in rather than the items and their usages. © 2020 Elsevier B.V. All rights reserved.Article Duygusal Konuşma Tanımada Yapay Veri Kullanımı(2025) Avcı, UmutBu çalışma, Türkçe konuşmalarda duygu tanıma performansını geliştirmek üzere veri artırma tekniklerinin rolünü incelemekte ve BUEMODB ile ITUDB veri kümelerini temel almaktadır. Konuşmaların sessiz bölümlerin kaldırılması ve ses sinyallerinin normalizasyonu ile gerçekleştirilen ön işleme aşamasının ardından, ses verileri mel spektrogramlara dönüştürülmüş, altı öznitelik seti çıkarılmış ve yedi farklı denetimli öğrenme algoritması kullanılarak temel sınıflandırma yapılmıştır. İlk deneyler sonucunda BUEMODB veri seti için %56,3, ITUDB veri seti için %65,2 F1 skoru elde edilmiştir. Sonraki deneylerde, veri artırma teknikleri kullanılarak eğitim verisi beş kat büyütülmüştür. Bu kapsamda Gürültü Ekleme ve Ses Tonu Değiştirme gibi ses dönüşümlerinin yanı sıra Yakınlaştırma ve Yükseklik Kaydırma gibi görüntü dönüşümleri uygulanmıştır. Ses bazlı tekniklerle veri artırıldığında sınıflandırma başarısı iyileşmiş, Hava Emilimi ve Zaman Ölçekleme kombinasyonu ile F1 skorları BUEMODB için %57,6’ya, ITUDB için %71,3’e çıkmıştır. Görüntü bazlı veri artırma teknikleri daha da yüksek performans göstererek BUEMODB için %60,0’lık, ITUDB için %73,2’lik F1 skorları sağlamıştır. Son olarak, en iyi sonuç veren ses ve görüntü dönüşümlerini birleştiren hibrit bir yaklaşım denenmiştir. Bu yöntemle BUEMODB için %59,7, ITUDB için %75,1 F1 skoruna ulaşılmış ve temel performansa göre yaklaşık %10’luk bir artış kaydedilmiştir. Bulgular, özellikle görüntü ve hibrit tabanlı veri artırma tekniklerinin dikkatlice seçilmesi halinde duygu tanıma doğruluğunun önemli ölçüde yükseltilebileceğini göstermiştir.Conference Object Citation - Scopus: 5Flight Gate Assignment Problem with Reinforcement Learning(Springer Science and Business Media Deutschland GmbH, 2023) Müge Muhafız Yıldız; Umut Avci; Mustafa Arslan Ornek; Cemalettin Öztürk; Muhafız Yıldız, Müge; Örnek, Mustafa Arslan; Öztürk, Cemalettin; Avcı, Umut; C. Kahraman , I.U. Sari , B. Oztaysi , S. Cevik Onar , S. Cebi , A.C. TolgaThe operation of an airport is a very complex task involving many actors. The primary mission of airport management is to provide sufficient capacity and the best working conditions to all airlines ground handling and service provider companies. Flight gate assignment is one of the essential planning problems airport management needs to address assigning incoming aircraft to the available gates or stands while satisfying operational constraints. Generally flight arrivals and departures are considered deterministic and various operational research methods have been applied to solve this combinatorial problem. However in real-life scenarios deterministic solutions are generally infeasible because arrival and departure times are uncertain. It is crucial to deal with these uncertainties to create a robust schedule. In this study we develop a Reinforcement Learning (RL) algorithm to solve the flight gate assignment problem since it is a sequential decision-making method and allows adaptive solutions to address urgent and frequent changes. © 2023 Elsevier B.V. All rights reserved.Conference Object Speech Emotion Recognition Using Spectrogram Patterns as Features(Springer Science and Business Media Deutschland GmbH info@springer-sbm.com, 2020) Umut Avci; Avci, Umut; A. Karpov , R. PotapovaIn this paper we tackle the problem of identifying emotions from speech by using features derived from spectrogram patterns. Towards this goal we create a spectrogram for each speech signal. Produced spectrograms are divided into non-overlapping partitions based on different frequency ranges. After performing a discretization operation on each partition we mine partition-specific patterns that discriminate an emotion from all other emotions. A classifier is then trained with features obtained from the extracted patterns. Our experimental evaluations indicate that the spectrogram-based patterns outperform the standard set of acoustic features. It is also shown that the results can further be improved with the increasing number of spectrogram partitions. © 2020 Elsevier B.V. All rights reserved.Conference Object Citation - Scopus: 6Handling Imbalanced Data in Predictive Maintenance: A Resampling-Based Approach(Institute of Electrical and Electronics Engineers Inc., 2023) Sejma Cicak; Umut Avci; Cicak, Sejma; Avci, UmutImbalanced data is a common problem in many areas and it can have significant impacts on the performance and generalizability of machine learning models. This is because the models fail to create a good representation of the examples in the minority class. This study aims at improving the classification success for the predictive maintenance tasks in which the data is generally imbalanced. To this end we use resampling methods that target creating balanced data. We present various oversampling and undersampling techniques and apply them to both synthetic and real-world datasets. We then perform classification experiments with imbalanced and balanced datasets by using different classifiers. The performances of different classifiers have been compared. More importantly we evaluate the effectiveness of resampling techniques to provide insights into their usefulness in handling class imbalance. Our study contributes to the growing body of literature on addressing the class imbalance in classification tasks and provides practical guidance for selecting appropriate sampling methods based on the characteristics of the dataset. © 2023 Elsevier B.V. All rights reserved.Article Citation - WoS: 2Citation - Scopus: 4A Comprehensive Analysis of Data Augmentation Methods for Speech Emotion Recognition(Institute of Electrical and Electronics Engineers Inc., 2025) Umut Avci; Avci, UmutThe limited availability of labeled emotional speech data remains a significant challenge in the development of robust speech emotion recognition systems. This paper presents a comprehensive investigation of the effectiveness of diverse data augmentation strategies for enhancing emotion recognition performance. Three different data augmentation categories were examined: audio-based transformations image-based modifications and feature-level synthesis. Seventeen transformations were used in audio-based data augmentation to change the time and frequency content of the raw audio signal. Eight transformations such as shifting rotating and zooming were applied to the spectrogram images for image-based data augmentation. The SpecAugment method was also used to transform the spectrograms into versions with masked time and frequency axes. In feature-space-based approaches new feature vectors were generated using five oversampling algorithms and a generative adversarial network. Experimental results from the EMO-DB and IEMOCAP datasets demonstrate that the data augmentation approaches enhance emotion classification performance by up to six percent. Empirical evidence indicates that training sets augmented through combinations of audio-based transformations yield the highest performance gains. In contrast the GAN-based approach fails to improve the classification performance. © 2025 Elsevier B.V. All rights reserved.
