Tuğrul K. Ateş

Content Based Copy Detection with Coarse Audio-Visual Fingerprints

Content Based Copy Detection (CBCD) emerges as a viable choice against active detection methodology of watermarking. The very first reason is that the media already under circulation cannot be marked and secondly, CBCD inherently can endure various severe attacks, which watermarking cannot. Although in general, media content is handled independently as visual and audio in this work both information sources are utilized in a unified framework, in which coarse representation of fundamental features are employed. From the copy detection perspective, number of attacks on audio content is limited with respect to visual case. Therefore audio, if present, is an indispensable part of a robust video copy detection system. In this study, the validity of this statement is presented through various experiments on a large data set.

A fast method for animated TV logo detection

As a recent trend some TV stations prefer to use animated logos, therefore the detection of the presence of an animated TV logo emerges as a new requirement for certain applications. In this paper we present a novel method for the detection of animated television logos in real-time. The main idea is to handle all frames of the animated logo in a unified manner. For this purpose a unified logo boundary representation is utilized. In the training stage, the boundaries of the animated logo from each frame are placed in a single set. During detection, a voting-based decision scheme is performed in order to determine the presence of the trained logo. Furthermore robustness of the method is improved by incorporating negative clues regarding the existence of the animated logo obtained from the region of interest. Aforementioned clues are unified in order to reach a final decision by using effective combination rules. Finally, time windowing is used for eliminating false positives with short durations. The proposed method is examined through typical broadcast data and promising results are obtained.

Adult Image Content Classification Using Global Features and Skin Region Detection

A method for adult content classification and nudity detection is presented. Objective of this method is to classify images into different classes, varying on the degree of adult content. We utilize MPEG-7 descriptors to represent visual information. Skin regions are detected to model adult content more precisely, as well as to eliminate false-positives. Proposed method is tested with conventional image sets. Experimental results indicate that the algorithm has an acceptable detection performance.

Speeding-up Pearson Correlation Coefficient calculation on graphical processing units

Sample correlation coefficient is used widely for finding signal similarity in data processing, multimedia, pattern recognition and artificial intelligence applications. Pearson Correlation Coefficient is the most common measure for the correlation coefficient between discrete signals. Similarity search in huge pattern databases require a fast way of calculating the correlation coefficient between numerical vectors. In this paper, a parallel and efficient way of calculating Pearson Correlation Coefficient on commodity central processing units (CPUs) and graphical processing units (GPUs) is proposed. Different implementations for C++, OpenCL and CUDA are compared over a vast number of architectures and through a wide parameter range. Experimental results are given in a comparative manner and investigated in both software and hardware perspectives.

Comparison of Quantization Index Modulation and Forbidden Zone Data Hiding for compressed domain video data hiding

Data hiding is now a part of daily life through various applications. In this work, we apply two data hiding methods, Quantization Index Modulation and Forbidden Zone Data Hiding, to video applications. We place these methods into a general video data hiding scheme and compare their performance against compression attacks. The results of the experiments with typical TV content indicate the superiority of FZDH, specifically for powerful attacks.

Multimodal concept detection in broadcast media: KavTan

Concept detection stands as an important problem for efficient indexing and retrieval in large video archives. In this work, the KavTan System, which performs high-level semantic classification in one of the largest TV archives of Turkey, is presented. In this system, concept detection is performed using generalized visual and audio concept detection modules that are supported by video text detection, audio keyword spotting and specialized audio-visual semantic detection components. The performance of the presented framework was assessed objectively over a wide range of semantic concepts (5 high-level, 14 visual, 9 audio, 2 supplementary) by using a significant amount of precisely labeled ground truth data. KavTan System achieves successful high-level concept detection performance in unconstrained TV broadcast by efficiently utilizing multimodal information that is systematically extracted from both spatial and temporal extent of multimedia data.

Relevance feedback for semantic classification: A comparative study

Immense increase in the number of multimedia content accessible from television and internet with the help developing technologies reveals efficient supervision and classification of such content as a problem. Relevance feedback is a technique which relies on evaluation of retrieval results by humans and enables reduce the semantic gap between ideas and low level representations. Content based high level classification system may employ relevance feedback for improved retrieval performance. In this paper, different relevance feedback algorithms, which can be utilized to increase generalized semantic classification performance, are discussed and compared inside an experimental framework. Some improvements are also proposed over obtained results.

Real-time arbitrary view rendering on GPU from stereo video and time-of-flight camera

Generating in-between images from multiple views of a scene is a crucial task for both computer vision and computer graphics fields. Photorealistic rendering, 3DTV and robot navigation are some of many applications which benefit from arbitrary view synthesis, if it is achieved in real-time. GPUs excel in achieving high computation power by processing arrays of data in parallel, which make them ideal for real-time computer vision applications. This paper proposes an arbitrary view rendering algorithm by using two high resolution color cameras along with a single low resolution time-of-flight depth camera and utilizing GPUs to achieve real-time processing rates. The presented ideas are examined in an experimental framework and based on the experimental results, it could be concluded that it is possible to realize content production and display stages of a free-viewpoint system in real-time by using only low-cost commodity computing devices.

Generalized visual concept detection

For efficient indexing and retrieval of video archives, concept detection stands as an important problem. In this work, a generalized structure that can be used for detection of diverse and distinct concepts is proposed. In the system, MPEG-7 Descriptors and Scale Invariant Transform (SIFT) are utilized as visual features. Furthermore, visual features are transformed by codebooks which are constructed by k-Means clustering. On the other hand, classification is performed on the distribution of visual features over the codebook. Proposed system is firstly tested against an elementary concept. Afterwards for a set of concepts system performance is reported on the TRECVID 2009 test set. It has been observed that with a sufficiently large training set high performance can be achieved with this method.

Multimodal concept detection on multimedia data- RTUK SKAAS KavTan system

Concept detection stands as an important problem for many applications like efficient indexing and retrieval in large video archives. In this work, for detection of diverse and distinct concepts a concept detection system (KavTan) that combines a variety of information sources under a single structure is proposed. The proposed system consists of Generalized Audio Concept Detection and Audio Keyword Detection sub-modules that use audio data and Generalized Visual Concept Detection, Video Text Detection, Human Detection, Nudity Detection, Blood Detection, Flag Detection and Skin Detection sub-modules that use visual data. Each concept is detected by using one or more of the mentioned modules. Proposed concept detection system is tested against multiple concepts and system performance is reported. It is observed that for most of the concepts high performance can be achieved with this approach.