arXiv subjects

Ahmet Emir Dirik

Publications and source records attributed to Ahmet Emir Dirik.

8 recordsLinked to original sources

arXiv2026

Evaluating SSL and ViViT Architectures for Cross-Corpus Audio MOS Prediction via LODO Validation

Automatic Mean Opinion Score (MOS) prediction is essential for evaluating large-scale synthetic speech and audio enhancement systems, yet models frequently struggle with domain shift. This study presents a comprehensive benchmarking of three architectural frameworks: Frozen Self-Supervised Learning (SSL-FRZ), Fine-Tuned SSL (SSL-FT), and a Video Vision Transformer (ViViT). Evaluation is conducted in two phases: Part I utilizes a consolidated corpus of 130,000 samples across 19 diverse datasets, while Part II focuses on a purified 17-dataset English-only corpus. To assess robustness, a systematic Leave-One-Dataset-Out (LODO) protocol is employed to quantify the generalization gap between seen and unseen distributions. Finally, the top-performing model is benchmarked against 18 state-of-the-art (SOTA) metrics using the ARECHO framework. Results demonstrate that an English-only purified corpus consistently yields higher predictive precision across all architectures. While SSL-FT achieves the highest performance on seen validation data, the SSL-FRZ model provides superior robustness on unseen distributions, achieving a competitive Mean Squared Error (MSE) of 0.36 on the URGENT 2024 benchmark-closely matching domain-optimized SOTA metrics (MSE 0.30). Although the ViViT architecture remains below SSL-based models in total capacity, it delivers stable results in English-only trials. LODO results confirm that while models perform significantly better on seen samples, frozen SSL embeddings combined with deep Transformer encoders offer the most stable and scalable solution for universal speech quality assessment. To support further research, the top-performing English-only SSL-Transformer model and weights are made publicly available via Hugging Face.

eess.AS↗

arXiv2025

Enhancing Diameter Measurement Accuracy in Machine Vision Applications

In camera measurement systems, specialized equipment such as telecentric lenses is often employed to measure parts with narrow tolerances. However, despite the use of such equipment, measurement errors can occur due to mechanical and software-related factors within the system. These errors are particularly evident in applications where parts of different diameters are measured using the same setup. This study proposes two innovative approaches to enhance measurement accuracy using multiple known reference parts: a conversion factor-based method and a pixel-based method. In the first approach, the conversion factor is estimated from known references to calculate the diameter (mm) of the unknown part. In the second approach, the diameter (mm) is directly estimated using pixel-based diameter information from the references. The experimental setup includes an industrial-grade camera and telecentric lenses. Tests conducted on glass samples (1-12 mm) and metal workpieces (3-24 mm) show that measurement errors, which originally ranged from 13-114 micrometers, were reduced to 1-2 micrometers using the proposed methods. By utilizing only a few known reference parts, the proposed approach enables high-accuracy measurement of all parts within the camera's field of view. Additionally, this method enhances the existing diameter measurement literature by significantly reducing error rates and improving measurement reliability.

cs.CV↗

arXiv2020

Fusion of Camera Model and Source Device Specific Forensic Methods for Improved Tamper Detection

PRNU based camera recognition method is widely studied in the image forensic literature. In recent years, CNN based camera model recognition methods have been developed. These two methods also provide solutions to tamper localization problem. In this paper, we propose their combination via a Neural Network to achieve better small-scale tamper detection performance. According to the results, the fusion method performs better than underlying methods even under high JPEG compression. For forgeries as small as 100$\times$100 pixel size, the proposed method outperforms the state-of-the-art, which validates the usefulness of fusion for localization of small-size image forgeries. We believe the proposed approach is feasible for any tamper-detection pipeline using the PRNU based methodology.

eess.IV↗

arXiv2019

PRNU Based Source Camera Attribution for Image Sets Anonymized with Patch-Match Algorithm

Patch-Match is an efficient algorithm used for structural image editing and available as a tool on popular commercial photo-editing software. The tool allows users to insert or remove objects from photos using information from similar scene content. Recently, a modified version of this algorithm was proposed as a counter-measure against Photo-Response Non-Uniformity (PRNU) based Source Camera Identification (SCI). The algorithm can provide anonymity at a great rate (97\%) and impede PRNU based SCI without the need of any other information, hence leaving no-known recourse for the PRNU-based SCI. In this paper, we propose a method to identify sources of the Patch-Match-applied images by using randomized subsets of images and the traditional PRNU based SCI methods. We evaluate the proposed method on two forensics scenarios in which an adversary makes use of the Patch-Match algorithm and distorts the PRNU noise pattern in the incriminating images he took with his camera. Our results show that it is possible to link sets of Patch-Match-applied images back to their source camera even in the presence of images that come from unknown cameras. To our best knowledge, the proposed method represents the very first counter-measure against the usage of of Patch-Match in the digital forensics literature.

eess.IV↗

arXiv2019

PRNU-Based Source Device Attribution for YouTube Videos

Photo Response Non-Uniformity (PRNU) is a camera imaging sensor imperfection which has earned a great interest for source device attribution of digital videos. A majority of recent researches about PRNU-based source device attribution for digital videos do not take into consideration the effects of video compression on the PRNU noise in video frames, but rather consider video frames as isolated images of equal importance. As a result, these methods perform poorly on re-compressed or low bit-rate videos. This paper proposes a novel method for PRNU fingerprint estimation from video frames taking into account the effects of video compression on the PRNU noise in these frames. With this method, we aim to determine whether two videos from unknown sources originate from the same device or not. Experimental results on a large set of videos show that the method we propose is more effective than existing frame-based methods that use either only I frames or all (I-B-P) frames, especially on YouTube videos.

eess.IV↗

arXiv2019

Detecting the Presence of ENF Signal in Digital Videos: a Superpixel based Approach

ENF (Electrical Network Frequency) instantaneously fluctuates around its nominal value (50/60 Hz) due to a continuous disparity between generated power and consumed power. Consequently, luminous intensity of a mains-powered light source varies depending on ENF fluctuations in the grid network. Variations in the luminance over time can be captured from video recordings and ENF can be estimated through content analysis of these recordings. In ENF based video forensics, it is critical to check whether a given video file is appropriate for this type of analysis. That is, if ENF signal is not present in a given video, it would be useless to apply ENF based forensic analysis. In this work, an ENF signal presence detection method is introduced for videos. The proposed method is based on multiple ENF signal estimations from steady superpixels, i.e. pixels that are most likely uniform in color, brightness, and texture, and intraclass similarity of the estimated signals. Subsequently, consistency among these estimates is then used to determine the presence or absence of an ENF signal in a given video. The proposed technique can operate on video clips as short as 2 minutes and is independent of the camera sensor type, i.e. CCD or CMOS.

cs.MM↗

arXiv2019

Analysis of Rolling Shutter Effect on ENF based Video Forensics

ENF is a time-varying signal of the frequency of mains electricity in a power grid. It continuously fluctuates around a nominal value (50/60 Hz) due to changes in supply and demand of power over time. Depending on these ENF variations, the luminous intensity of a mains-powered light source also fluctuates. These fluctuations in luminance can be captured by video recordings. Accordingly, ENF can be estimated from such videos by analysis of steady content in the video scene. When videos are captured by using a rolling shutter sampling mechanism, as is done mostly with CMOS cameras, there is an idle period between successive frames. Consequently, a number of illumination samples of the scene are effectively lost due to the idle period. These missing samples affect ENF estimation, in the sense of the frequency shift caused and the power attenuation that results. This work develops an analytical model for videos captured using a rolling shutter mechanism. The model illustrates how the frequency of the main ENF harmonic varies depending on the idle period length, and how the power of the captured ENF attenuates as idle period increases. Based on this, a novel idle period estimation method for potential use in camera forensics that is able to operate independently of video frame rate is proposed. Finally, a novel time-of-recording verification approach based on use of multiple ENF components, idle period assumptions and interpolation of missing ENF samples is also proposed.

cs.MM↗

arXiv2019

Deep Learning Based Vehicle Make-Model Classification

This paper studies the problems of vehicle make & model classification. Some of the main challenges are reaching high classification accuracy and reducing the annotation time of the images. To address these problems, we have created a fine-grained database using online vehicle marketplaces of Turkey. A pipeline is proposed to combine an SSD (Single Shot Multibox Detector) model with a CNN (Convolutional Neural Network) model to train on the database. In the pipeline, we first detect the vehicles by following an algorithm which reduces the time for annotation. Then, we feed them into the CNN model. It is reached approximately 4% better classification accuracy result than using a conventional CNN model. Next, we propose to use the detected vehicles as ground truth bounding box (GTBB) of the images and feed them into an SSD model in another pipeline. At this stage, it is reached reasonable classification accuracy result without using perfectly shaped GTBB. Lastly, an application is implemented in a use case by using our proposed pipelines. It detects the unauthorized vehicles by comparing their license plate numbers and make & models. It is assumed that license plates are readable.

cs.CV↗