Search

Estimation of spatial-temporal transformation consistency in object level between the original video and its copies is one of the main issues arise in video fingerprinting. The key purpose of this research is to discover an efficient and excellent approach for the above mentioned problematic situation even though the comparison happens with longer length videos under various types of photometric and geometric transformations. Convert classical copy detection into high level pattern matching task was the proposed and applied approach in this scenario. In here frame wise SIFT interest point trajectory data provide precise and invariant information of relevant video in pattern matching phase. Finally classical challenges of video copy detection like brightness changes, blurs, zooming, size changes, cropping, illumination changes, noise, ratio changes, etc. were addressed by proposed methodology successfully. [ABSTRACT FROM AUTHOR]

This paper tackles the issue of retrieving different instances of an object of interest within a given video document or in a video database. The principle consists of considering a semi-global image representation based on an over-segmentation of video frames. An aggregation mechanism is then applied in order to group a set of segments into an object similar to the query, under a global similarity criterion. We test the effectiveness of three different aggregation strategies, two of them based on a greedy approach, and the third one involving simulated annealing optimization. Experimental results on different color spaces show promising performances, with First Tier and Bull Eye detection rates of up to 70% and 88%, respectively. The integration of the method in a web-based video navigation system, allowing fast video object retrieval, is finally described. [ABSTRACT FROM PUBLISHER]

In the context of content-based indexing applications, the automatic classification and interpretation of visual content is a key issue that needs to be solved. This paper proposes a novel approach for semantic video object interpretation. The principle consists of exploiting the a priori information contained in categorized 3D model data sets, in order to transfer the semantic labels from such models to unknown video objects. Each 3D model is represented as a set of 2D views, described with the help of shape descriptors. A matching technique is used in order to perform an association between categorized 3D models and 2D video objects. The experimental evaluation shows the interest of our approach, which yields recognition rates of up to 92.5%. [ABSTRACT FROM PUBLISHER]

No longer is video creation and storage solely in the hands of professionals. Video repositories are growing at an astounding rate due advances in multimedia technologies. The vast size of video repositories presents challenges for users attempting to identify preferred content. Automated methods for content discovery are necessary to meet the needs of users. One of the more challenging areas of video content discovery is in identifying affective, or emotional, video content. Automatic affective video indexing techniques attempt to use computer-based methods to automatically identify content in videos that is affective in nature. This is the first known automatic affective video indexing study that focuses on slapstick, one of the most popular types of humor techniques. The study shows positive results and contributes to the field by identifying the targeted affective content without relying on actual human emotional responses. [ABSTRACT FROM PUBLISHER]

This paper proposes an original method for video indexing based on a spatio-temporal segmentation scheme. The basic idea is to extract salient regions from the video content and use them as scene descriptors for indexing. The obtained results confirm the efficiency of the proposed approach and open new perspectives for video summarizing and indexing. [ABSTRACT FROM PUBLISHER]

This paper describes a method that extracts the start of every real golf swing in golf practice video recorded from face-on camera angle for video indexing. This new method accelerates and improves the process of golf swing analysis by allowing golfer to quickly navigate to the starting point of each real swing. Real swing is the most important part of golf practice video because it truly shows the problem of golfer's swing. Each start of real swing is extracted by detecting takeaway action at the start of the swing. The system consists of five main processes which are golf club head detection, golf ball detection, takeaway detection, takeaway validation and hit detection. First the golf club shaft is detected by using frame difference, morphological operation, canny edge detection and Hough line transform respectively. After that golf club head is extracted and in order to detect takeaway action accurately, the system also detects golf balls position using HSV color space to extract near white color. Then the information of detected golf clubs and golf balls are used to identify the start of each real swing. The result of experiments on golf practice video with different camera setups and light conditions shows that this method can efficiently extract the beginning of real swings. [ABSTRACT FROM PUBLISHER]

The segmentation of road images from vehicle mounted video is a challenging and difficult problem. One of the problems is the presence of different types of objects and not all objects are present in the same frame. For example, road sign is not visible in all frames. In this paper, we propose a novel framework for segmenting road images in a hierarchical manner that can separate the following objects: sky, road, road signs, and vegetation from the video data. Each frame in the video is analysed separately. The hierarchical approach does not assume the presence of a certain number of objects in a single frame. We have also developed a segmentation framework based on SVM learning. The proposed framework has been tested on the Transport and Main Roads Queensland's video data. The experimental results indicate that the proposed framework can detect different objects with an accuracy of 95.65%. [ABSTRACT FROM PUBLISHER]

This paper covers the application of corpus-based and corpus-driven methods to the research of collocation in modern Chinese, summarizing and commenting on the basic concepts, terms, formulae, approaches and processes in collocation research. Based on previous studies, this paper raises suggestions in terms of the method of extracting collocates, determination of span, colligation and semantic prosody, and emphasizes on the use of statistics, manual proofreading and the combination of quantitative and qualitative methods in collocation study based on the characteristics of collocation in Chinese. [ABSTRACT FROM PUBLISHER]

Lecture videos have been commonly used to supplement in-class teaching and for distance learning. Videos recorded during in-class teaching and made accessible online are a versatile resource on par with a textbook and the classroom itself. Nonetheless, the adoption of lecture videos has been limited, in large part due to the difficulty of quickly accessing the content of interest in a long video lecture. In this work, we present “video indexing” and “keyword search” that facilitate access to video content and enhances user experience. Video indexing divides a video lecture into segments indicating different topics by identifying scene changes based on the analysis of the difference image from a pair of video frames. We propose an efficient indexing algorithm that leverages the unique features of lecture videos. Binary search with frame sampling is employed to efficiently analyze long videos. Keyword search identifies video segments that match a particular keyword. Since text in a video frame often contains a diversity of colors, font sizes and backgrounds, our text detection approach requires specialized preprocessing followed by the use of off-the-shelf OCR engines, which are designed primarily for scanned documents. We present image enhancements: text segmentation and inversion, to increase detection accuracy of OCR tools. Experimental results on a suite of diverse video lectures were used to validate the methods developed in this work. Average processing time for a one-hour lecture is around 14 minutes on a typical desktop. Search accuracy of three distinct OCR engines - Tesseract, GOCR and MODI increased significantly with our preprocessing transformations, yielding an overall combined accuracy of 97%. The work presented here is part of a video streaming framework deployed at multiple campuses serving hundreds of lecture videos. [ABSTRACT FROM PUBLISHER]

Video-equipped mini unmanned aerial vehicles (mini-UAVs) are becoming increasingly popular for surveillance, remote sensing, law enforcement, and search and rescue operations, all of which rely on thorough coverage of a target observation area. However, coverage is not simply a matter of seeing the area (visibility) but of seeing it well enough to allow detection of targets of interest, a quality we here call "see-ability". Video flashlights, mosaics, or other geospatial compositions of the video may help place the video in context and convey that an area was observed, but not necessarily how well or how often. This paper presents a method for using UAV-acquired video georegistered to terrain and aerial reference imagery to create geospatial video coverage quality maps and indices that indicate relative video quality based on detection factors such as image resolution, number of observations, and variety of viewing angles. When used for offline post-analysis of the video, or for online review, these maps also enable geospatial quality-filtered or prioritized non-sequential access to the video. We present examples of static and dynamic see-ability coverage maps in wilderness search-and-rescue scenarios, along with examples of prioritized non-sequential video access. We also present the results of a user study demonstrating the correlation between see-ability computation and human detection performance. [ABSTRACT FROM AUTHOR]

Adaptive assessment of video based m-training is an area which involves many technologies. This paper presents the design and implementation of an m-Learning system that has video indexing and automatic caption creation features. The system is able to adjust the m-Learning content based on the progress of the participants. Unlike most mobile devices, in which only downloading is enabled, the system is able to not only download the content to the devices, but also upload data such as learning results, answers to quizzes, etc. back to the remote databases. It provides a two-way interactive learning experience. In addition, the system is capable of creating video quizzes and searching videos using text from the speech. The research also focuses on collaborative learning, video indexing and automatic caption creation. The m-Learning a modules, together with other resources construct a one-point interface for people to access by both PDAs and PCs. This system is being tested by educational professionals on an online system with monthly traffic of over 30,000 hits. [ABSTRACT FROM AUTHOR]