Yazar "Hanilci, Cemal" seçeneğine göre listele
Listeleniyor 1 - 10 / 10
Sayfa Başına Sonuç
Sıralama seçenekleri
Öğe Benchmarking domain adaptation for LiDAR-based 3D object detection in autonomous driving(Springer London Ltd, 2025) Balim, Mustafa Alper; Hanilci, Cemal; Acir, NurettinThe generalization capability of 3D object detection models is crucial for ensuring robust perception in autonomous driving systems. While state-of-the-art models such as Voxel R-CNN, PV-RCNN, and CenterPoint have demonstrated strong performance on publicly available datasets (e.g., KITTI, Waymo, and nuScenes). In this study, we conduct a comprehensive benchmark evaluation. We introduce two custom datasets: (i) a real-world dataset collected using KARSAN's autonomous minibus equipped with a 128-channel LiDAR sensor under diverse traffic conditions, and (ii) a simulated dataset generated using the AWSIM simulation platform, capturing over five hours of synthetic driving data with virtual LiDAR sensors. Our results indicate that 3D object detection performance is highly dataset-dependent, as no single model achieves superior results across all datasets and metrics. Cross-dataset evaluation highlights the challenges of domain mismatch, which causes significant performance degradation when models are tested on our custom datasets, particularly in the synthetic domain. To mitigate these effects, we explore six domain adaptation techniques and demonstrate that their application substantially improves model performance. Bi3D, SESS, and Uni3D outperform UDA, CLUE, and ST3D, yielding more robust generalization across both real-world and simulated environments. These findings shed light on the potential of domain adaptation to improve model performance across domain shifts, despite the ongoing challenges in achieving consistent outcomes across all environments.Öğe Comparative Analysis of Audio Steganography Methods(2022) Beyin, Funda Aslantaş; Hanilci, CemalInformation security has been one of the most important issues of all time for both individuals and companies. Delivering data to the correct recipient is crucial for personal data protection, the privacy of personal life, and national security. To this end, different methods have been developed over the years to hide information from malicious individuals. Steganography is one of the most important information hiding methods that received great attention. In this study, five different audio steganography techniques (least significant bit, echo hiding, wavelet coding, spread spectrum, and cepstrum) are utilized and a comparison of these techniques is performed on Turkish audio recordings.To this end, hidden messages of various sizes were embedded into 20 audio recordings from 10 male and 10 female speakers using different embedding algorithms. Signal-to-noise ratios (SNR) computed between stego and cover audio files show that embedded message length and frame size are the main factors that determine the quality. In addition, it is observed that there is no perceptual difference between the cover and stego audio recordings. Hence, the human auditory system is unable to determine whether an audio recording is authentic or conveys a hidden message. Experimental results show that as the message length increases, the average SNR value decreases irrespective of the steganography technique, as expected. The well-known least significant bit (LSB) technique yields the highest average SNR value among the five steganography methods. The spectrographic comparison of the cover and stego audio recordings shows that hiding the secret message in an original audio signal highly affects the high-frequency region more than the low-frequency components.Öğe Double Compressed Wideband AMR Speech Detection Using Deep Neural Networks(Springer Birkhauser, 2024) Buker, Aykut; Hanilci, CemalDetecting double compressed (DC) speech signals is an important audio forensics task since it is highly related to the integrity and the authenticity of the recording. Adaptive multi-rate (AMR) speech codec is a popular audio compression technique specifically optimized for speech signals and it is a standard audio recording format in the vast majority of the smart phones. All of the previous studies addressing the detection of DC AMR signals report their findings for the speech signals compressed using the narrowband AMR codec (AMR-NB). Meanwhile, wideband AMR codec (AMR-WB) has been used by several mobile phone manufacturers, but DC AMR-WB speech signal detection performance remains unknown. To the best of our knowledge, this is the first study focusing on detecting the DC signals compressed using the AMR-WB speech codec. To this end, we propose three different deep neural network-based DC AMR-WB signal detection systems where the spectrogram representations of the speech signals are used as the input features. Experimental results conducted on TIMIT database provide several important findings regarding the DC AMR-WB speech detection. Firstly, DC AMR-WB detection is found to be a more challenging task than detecting the AMR-NB signals. For example, convolutional neural network (CNN)-based system yields 74.83% and 99.93% detection rates on AMR-WB and AMR-NB coded signals, respectively. Secondly, capturing the temporal information using long short-term memory (LSTM) network with the DC AMR-WB signal detection accuracy of 86.25% is found to be superior to the CNN system. Thirdly, combining the deep feature representations learned by CNN and LSTM networks further improves the performance. Fourthly, the detection rates are found to deteriorate when the signals are first encoded using different audio codecs prior to AMR-WB compression. Finally, applying score level or decision level fusion to the proposed three systems improves the detection rates, in general.Öğe Evaluating Parameter Sharing for Spoofing-Aware Speaker Verification: A Case Study on the ASVspoof 5 Dataset(Isca-Int Speech Communication Assoc, 2025) Buker, Aykut; Kurnaz, Oguzhan; Bekiryazici, Yule; Demirtac, Selim Can; Hanilci, CemalSpoofing-aware speaker verification (SASV) is an important but challenging task and has been a primary focus of the recently organized ASVspoof 5 challenge. As SASV integrates automatic speaker verification (ASV) and countermeasure (CM) systems, its performance depends on the effectiveness of each system. This study systematically examines the impact of different parameter-sharing (PS) strategies, which facilitate joint optimization, on SASV performance using the ASVspoof 5 dataset. Experimental results indicate that PS enhances performance for specific attack types and codec conditions. For example, the baseline system achieves a min a-DCF of 0.329 on the A26 attack, which improves to 0.233 with PS. Similarly, for AMR-compressed signals, PS yields a 14.09% performance gain. These observations show that PS techniques are effective in mitigating certain spoofing attacks and improving robustness to degraded audio conditions in SASV systems.Öğe Exploring the Effectiveness of the Phase Features on Double Compressed AMR Speech Detection(Mdpi, 2024) Buker, Aykut; Hanilci, CemalDetermining whether an audio signal is single compressed (SC) or double compressed (DC) is a crucial task in audio forensics, as it is closely linked to the integrity of the recording. In this paper, we propose the utilization of phase spectrum-based features for detecting DC narrowband and wideband adaptive multi-rate (AMR-NB and AMR-WB) speech. To the best of our knowledge, phase spectrum features have not been previously explored for DC audio detection. In addition to introducing phase spectrum features, we propose a novel parallel LSTM system that simultaneously learns the most representative features from both the magnitude and phase spectrum of the speech signal and integrates both sets of information to further enhance its performance. Analyses demonstrate significant differences between the phase spectra of SC and DC speech signals, suggesting their potential as representative features for DC AMR speech detection. The proposed phase spectrum features are found to perform as well as magnitude spectrum features for the AMR-NB codec, while outperforming the magnitude spectrum in detecting AMR-WB speech. The proposed phase spectrum features yield 8% performance improvement in terms of true positive rate over the magnitude spectrogram features. The proposed parallel LSTM system further improves DC AMR-WB speech detection.Öğe Image Forgery Detection Based On Parallel Convolutional Neural Networks(Ieee, 2022) Korkmaz, Ahmet; Hanilci, CemalAs a result of the advancement of software tools used in digital image processing, it has become very easy to generate fake images by applying various manipulations techniques on the original (authentic) images. These manipulated images can easily be used with malicious intentions in important fields such as law, medicine and communication. Hence, image forgery detection, determining whether an image is original or forged, is an important task. In this study, an image forgery detection system is proposed by combining three deep neural network structures in parallel, unlike the uniform deep learning methods used in image forgery detection. The proposed method has been evaluated on three different datasets, and the results clearly demonstrate the efficiency of the proposed method with promising classification accuracy.Öğe Investigating the Potential of Multi-Stage Score Fusion in Spoofing-Aware Speaker Verification(Ieee, 2025) Kurnaz, Oguzhan; Kinnunen, Tomi H.; Hanilci, CemalDespite improvements in automatic speaker verification (ASV), vulnerability against spoofing attacks remains a major concern. In this study, we investigate the integration of ASV and countermeasure (CM) subsystems into a modular spoof-aware speaker verification (SASV) framework. Unlike conventional single-stage score-level fusion methods, we explore the potential of a multi-stage approach that utilizes the ASV and CM systems in multiple stages. By leveraging ECAPA-TDNN (ASV) and AASIST (CM) subsystems, we consider support vector machine and logistic regression classifiers to achieve SASV. In the second stage, we integrate their outputs with the original score to revise fusion back-end classifiers. Additionally, we incorporate another auxiliary score from RawGAT (CM) to further enhance our SASV framework. Our approach yields an equal error rate (EER) of 1.30% on the evaluation dataset of the SASV2022 challenge, representing a 24% relative improvement over the baseline system.Öğe Optimizing a-DCF for Spoofing-Robust Speaker Verification(Ieee-Inst Electrical Electronics Engineers Inc, 2025) Kurnaz, Oguzhan; Mishra, Jagabandhu; Kinnunen, Tomi H.; Hanilci, CemalAutomatic speaker verification (ASV) systems are vulnerable to spoofing attacks. We propose a spoofing-robust ASV system optimized directly for the recently introduced architecture-agnostic detection cost function (a-DCF), which allows targeting a desired trade-off between the contradicting aims of user convenience and robustness to spoofing. We combine a-DCF and binary cross-entropy (BCE) with a novel straightforward threshold optimization technique. Our results with an embedding fusion system on ASVspoof2019 data demonstrate relative improvement of 13% over a system trained using BCE only (from minimum a-DCF of 0.1445 to 0.1254). Using an alternative non-linear score fusion approach provides relative improvement of 43% (from minimum a-DCF of 0.0508 to 0.0289).Öğe Tekrar oynatma saldırılarının tespitinde genlik ve faz spektrumu tabanlı özniteliklerin derin öğrenme ile karşılaştırmalı analizi(2025) Bekiryazıcı, Şule; Hanilci, Cemal; Ozcan, NeyirGelişen dijital teknolojiler ve artan güvenlik ihtiyaçları, biyometrik kimlik doğrulama sistemlerine olan ilgiyi ciddi biçimde artırmıştır. Bu sistemler, bireylerin fiziksel ya da davranışsal özelliklerini temel alarak yüksek güvenlik düzeyinde kimlik doğrulama sağlamaktadır. Otomatik konuşmacı doğrulama (Automatic Speaker Verification- ASV) sistemleri, kullanıcı dostu yapıları ve doğal etkileşim avantajlarıyla öne çıkmaktadır. Ancak bu sistemler, özellikle tekrar oynatma (replay) saldırıları gibi yapay giriş yöntemlerine karşı savunmasız kalabilmektedir. Bu tür saldırılar, önceden kaydedilen bir konuşma örneğinin yeniden çalınarak sistemin aldatılması temeline dayanmakta ve düşük maliyetli, uygulanabilir yöntemler olmaları nedeniyle önemli bir tehdit oluşturmaktadır. Bu çalışmada, konuşma sinyallerinden türetilen genlik ve faz spektrumu temelli özniteliklerin, tekrar oynatma saldırılarının tespiti üzerindeki etkileri sistematik olarak incelenmiştir. Çalışma kapsamında sekiz farklı genlik temelli ve üç farklı faz temelli öznitelik çıkarılmış, bu öznitelikler ASVspoof-2017, ASVspoof-2019 (Physical Access) ve ASVspoof-2021 (Physical Access) veri kümeleri üzerinde test edilmiştir. Her bir öznitelik kümesi, konuşma sinyalinin spektral veya fazsal özelliklerini farklı yönleriyle temsil etmekte; özellikle yankı, bozulma ve yeniden kaydetme gibi saldırı izlerini ortaya çıkarma potansiyeline sahiptir. Derin öğrenme temelli sınıflandırma için ResNet ve LCNN mimarileri kullanılmış, sistem başarımı EER (Equal Error Rate) ve t-DCF (tandem Detection Cost Function) metrikleri ile değerlendirilmiştir. Özellikle yüksek frekans bantlarında elde edilen ayırt edici spektral özelliklerin, sahte konuşma sinyallerinin tespitinde önemli bir katkı sunduğu gözlemlenmiştir. Elde edilen bulgular hem öznitelik çeşitliliği hem de mimari karşılaştırmaları açısından literatüre yeni ve bütüncül bir perspektif kazandırmaktadır.Öğe Toward robust replay attack detection in Automatic Speaker Verification: A study of spectrum estimation and channel magnitude response modeling(Academic Press Ltd- Elsevier Science Ltd, 2026) Bekiryazici, Sule; Hanilci, Cemal; Ozcan, NeyirAutomatic Speaker Verification (ASV) systems are increasingly adopted for biometric authentication but remain highly vulnerable to spoofing, particularly replay attacks. Existing countermeasures (CMs) for replay attack detection rely predominantly on discrete Fourier transform (DFT)-based spectral features, which are sensitive to noise and channel distortions common in physical access (PA) scenarios. This work presents the first comprehensive study of Channel Magnitude Response (CMR) representations for replay detection, explicitly analyzing the impact of spectrum estimation and feature design. The contribution of this work are fourfold: (i) CMR estimation is generalized beyond MFCCs to LFCC and CQCC features, with LFCC-based CMRs offering superior discrimination; (ii) alternative spectrum estimators - linear prediction (LP) and multitaper (MT) - are integrated into the CMR pipeline, yielding substantial gains over conventional DFT (iii) robustness is investigated under silence-free (voiced-only) conditions, mitigating known biases in ASVspoof datasets and (iv) a systematic evaluation of CMR is provided on the recently released ReplayDF corpus, a challenging benchmark combining replay and synthetic speech variability. Experiments on ASVspoof 2017, 2019, 2021, and ReplayDF using both baseline classifiers (ResNet18 and LCNN) and stronger models (Res2Net50 and SE-Res2Net50) show that the proposed approach consistently outperforms conventional features. Particularly, LFCC-CMR features with LP spectra achieve an Equal Error Rate (EER) as low as 1.34% on ASVspoof 2019 (PA), representing considerable relative improvements over traditional methods. Moreover, CMR-based systems retain high performance even when silent segments are removed, unlike conventional approaches. These results establish CMR with principled spectral modeling as a robust and generalizable framework for replay attack detection, opening new directions for resilient spoofing countermeasures.
