Bilgisayar Mühendisliği Bölümü Yayın Koleksiyonu
Permanent URI for this collectionhttps://hdl.handle.net/20.500.12416/253
Browse
5 results
Search Results
Article Citation - WoS: 3Citation - Scopus: 5Block Size Analysis for Discrete Wavelet Watermarking and Embedding a Vector Image as a Watermark(Zarka Private Univ, 2019) Sever, Hayri; Sever, Hayri; Senol, Ahmet; Elbasi, Ersin; Bilgisayar MühendisliğiAs telecommunication and computer technologies proliferate, most data are stored and transferred in digital format. Content owners, therefore, are searching for new technologies to protect copyrighted products in digital form. Image watermarking emerged as a technique for protecting image copyrights. Early studies on image watermarking used the pixel domain whereas modern watermarking methods convert a pixel based image to another domain and embed a watermark in the transform domain. This study aims to use, Block Discrete Wavelet Transform (BDWT) as the transform domain for embedding and extracting watermarks. This study consists of 2 parts. The first part investigates the effect of dividing an image into non overlapping blocks and transforming each image block to a DWT domain, independently. Then, effect of block size on watermark success and, how it is related to block size, are analyzed. The second part investigates embedding a vector image logo as a watermark. Vector images consist of geometric objects such as lines, circles and splines. Unlike pixel-based images, vector images do not lose quality due to scaling. Vector watermarks deteriorate very easily if the watermarked image is processed, such as compression or filtering. Special care must be taken when the embedded watermark is a vector image, such as adjusting the watermark strength or distributing the watermark data into the image. The relative importance of watermark data must be taken into account. To the best of our knowledge this study is the first to use a vector image as a watermark embedded in a host image.Article Citation - WoS: 5Citation - Scopus: 11A Shallow 3d Convolutional Neural Network for Violence Detection in Videos(Cairo Univ, Fac Computers & information, 2024) Kaya, Aydin; Sever, Hayri; Dundar, Naz; Keceli, Ali SeydiWith the recent worldwide statistical rise in the amount of public violence, automated violence detection in surveillance cameras has become a matter of high importance. This work introduces an end-to-end, trainable 3D Convolutional Neural Network (3D CNN) for detecting violence in video footage. The proposed network is inherently capable of processing both spatial and temporal information, thereby obviating the need for additional models that would introduce higher computational requirements and complexity. This work has two main contributions: 1) developing a lightweight 3D CNN suitable for inference on edge devices as mobile systems, and 2) a comprehensive explanation of all components comprising a CNN model, thereby enhances model interpretability. Experiments were conducted to assess the performance of the proposed model using a consolidated dataset combining four benchmark datasets. The results of the experiments support the asserted contributions, which are discussed in detail.Article Citation - WoS: 3Citation - Scopus: 3Binary Background Model With Geometric Mean for Author-Independent Authorship Verification(Sage Publications Ltd, 2023) Sezer, Ebru A.; Sever, Hayri; Canbay, PelinAuthorship verification (AV) is one of the main problems of authorship analysis and digital text forensics. The classical AV problem is to decide whether or not a particular author wrote the document in question. However, if there is one and relatively short document as the author's known document, the verification problem becomes more difficult than the classical AV and needs a generalised solution. Regarding to decide AV of the given two unlabeled documents (2D-AV), we proposed a system that provides an author-independent solution with the help of a Binary Background Model (BBM). The BBM is a supervised model that provides an informative background to distinguish document pairs written by the same or different authors. To evaluate the document pairs in one representation, we also proposed a new, simple and efficient document combination method based on the geometric mean of the stylometric features. We tested the performance of the proposed system for both author-dependent and author-independent AV cases. In addition, we introduced a new, well-defined, manually labelled Turkish blog corpus to be used in subsequent studies about authorship analysis. Using a publicly available English blog corpus for generating the BBM, the proposed system demonstrated an accuracy of over 90% from both trained and unseen authors' test sets. Furthermore, the proposed combination method and the system using the BBM with the English blog corpus were also evaluated with other genres, which were used in the international PAN AV competitions, and achieved promising results.Conference Object Citation - WoS: 2Citation - Scopus: 2Clinical Decision Support Systems: From the Perspective of Small and Imbalanced Data Set(Ios Press, 2019) Akcapinar Sezer, Ebru; Sever, Hayri; Par, Oznur EsraClinical decision support systems are data analysis software that supports health professionals' decision - making the process to reach their ultimate outcome, taking into account patient information. However, the need for decision support systems cannot be denied because of most activities in the field of health care within the decision-making process. Decision support systems used for diagnosis are designed based on disease due to the complexity of diseases, symptoms, and disease-symptoms relationships. In the design and implementation of clinical decision support systems, mathematical modeling, pattern recognition and statistical analysis techniques of large databases and data mining techniques such as classification are also widely used. Classification of data is difficult in case of the small and / or imbalanced data set and this problem directly affects the classification performance. Small and/or imbalance dataset has become a major problem in data mining because classification algorithms are developed based on the assumption that the data sets are balanced and large enough. Most of the algorithms ignore or misclassify examples of the minority class, focus on the majority class. Most health data are small and imbalanced by nature. Learning from imbalanced and small data sets is an important and unsettled problem. Within the scope of the study, the publicly accessible data set, hepatitis was oversampled by distance-based data generation methods. The oversampled data sets were classified by using four different machine learning algorithms. Considering the classification scores of four different machine learning algorithms (Artificial Neural Networks, Support Vector Machines, Naive Bayes and Decision Tree), optimal synthetic data generation rate is recommended.Article Citation - WoS: 7Citation - Scopus: 10Extending a Sentiment Lexicon With Synonym-Antonym Datasets: Swnettr Plus(Tubitak Scientific & Technological Research Council Turkey, 2019) Genc, Burkay; Sever, Hayri; Saglam, FatihIn our previous studies on developing a general-purpose Turkish sentiment lexicon, we constructed SWNetTR-PLUS, a sentiment lexicon of 37K words. In this paper, we show how to use Turkish synonym and antonym word pairs to extend SWNetTR-PLUS by almost 33% to obtain SWNetTR++, a Turkish sentiment lexicon of 49K words. The extension was done by transferring the problem into the graph domain, where nodes are words, and edges are synonym- antonym relations between words, and propagating the existing tone and polarity scores to the newly added words using an algorithm we have developed. We tested the existing and new lexicons using a manually labeled Turkish news media corpus of 500 news texts. The results show that our method yielded a significantly more accurate lexicon than SWNetTR-PLUS, resulting in an accuracy increase from 72.2% to 80.4%. At this level, we have now maximized the accuracy rates of translation-based sentiment analysis approaches, which first translate a Turkish text to English and then do the analysis using English sentiment lexicons.