Jongho Nang

A video streaming system for mobile phones: practice and experience

This paper presents a case of video streaming system for mobile phone which has actually been implemented and deployed for commercial services in CDMA2000 1X cellular phone networks. As the computing environment and the network connection of cellular phones are significantly different from the wired desktop environment, the traditional desktop streaming method is not applicable. Therefore, a new architecture is required to suit the successfully streaming in the mobile phone environment. We have developed a very lightweight video player for use in mobile phone and the related authoring tool for the player. The streaming server has carefully been designed to provide high efficiency, reliability and scalability. Based on a specifically-designed suite of streaming protocol, the server employs an adaptive rate control mechanism which transmits the media packets appropriately into the network according to the change in network bandwidth.

Energy management for interactive applications in mobile handheld systems

The usage of interactive applications increases in handheld systems. In this paper, we describe a system-level dynamic power management scheme that considers interaction between the CPU and the WNIC, and interactive applications to reduce the energy consumption of handheld systems. Previous research efforts considered the CPU and the WNIC separately to reduce energy consumption. The proposed scheme reduces the energy consumption of handheld systems by using the information gathered from the WNIC to control the CPU voltage and frequency when interactive applications are executed. Experimental results show that on average the proposed scheme reduces energy consumption by 46% when compared to DVFS (Dynamic Voltage and Frequency Scaling) for the CPU and DPM (Dynamic Power Management) for the WNIC.

A Novel Approach to Collect Training Images from WWW for Image Thesaurus Building

This paper introduces a novel approach to change gathered images from WWW into training images to build an image thesaurus. The requirements for being training images are a large number of images and with highly relevant to a given concept. To fulfil these requirements, a system should be able to collect a large number of relevant images to a given concept from WWW by the proposed criterion of relevance to the concept for each image. Then, the irrelevant images would be filtered out by the modified hierarchical clustering method based on the weighted combination of 5 MPEG-7 visual descriptors and the proposed criterion of relevance to the concept for each cluster. Upon experimental results, the precision of the set of images generated by the proposed method is about 18% higher than that of the set of images generated by other methods

Web Image Annotation Based on the Decision Rules Inferred by the Statistical Analysis of Web Pages

This paper proposes a rule based web image annotation method which improves the precision and recall of annotation by the use of decision tree. This decision tree learns the relationship between images and their annotations based on the proposed 17 attributes that specify the structural relationship between them in HTML documents and the visual characteristics of the images. By converting and pruning this learned tree, a set of rules with high estimated accuracy which determines whether or not a word can be the keyword of an image can be generated. Upon experimental results, the proposed method made 57 rules and the precision and recall of annotation by these rules were about 88% and 95% for the various concepts, respectively. We argue the contribution of this work in two aspects. First, we suggest the clear criteria for precise annotation inferred by the statistical analysis of many web pages. Second, to cope with the deterioration of recall caused by the lack of measure for the visual characteristics, the visual similarity between an image and its concept combines to the attributes that used for tree learning.

Analyzing Performance of MPEG-7 Video Signature for Video Copy Detection

In recent years, we can access to video contents anywhere and at any time. Therefore distributed video is easily copied, transformed and republished. Since it brings copyright problem, similarity/duplicate detection and measurement is essential to identify the excessive content duplication. In this paper, we analysed various discernment of video which has been transformed with various ways using MPEG-7 Video Signature. MPEG-7 Video Signature, one of video copy detection algorithms, is block based abstraction. Thus we assume Video Signature is weak for spatial transform. The experiments show that MPEG-7 Video Signature is very weak for spatial transform which could occur general as we have assumed.

Content based web image retrieval system using both MPEG-7 visual descriptors and textual information

This paper introduces a complete content based web image retrieval system by which images on WWW are automatically collected, searched and browsed using both visual and textual features. To improve the quality of search results and the speed of retrieval, we propose two new algorithms such as a keyword selection algorithm using visual features as well as the layout of web page, and a k-NN search algorithm based on the hierarchical bitmap index [17] using multiple features with dynamically updated weights. Moreover, these algorithms are adjusted for the MPEG-7 visual descriptors [14] that are used to represent the visual features of image in our system. Experimental results of keyword selection and image retrieval show the superiority of proposed algorithms and a couple of visual interfaces of the system are presented to help understanding some retrieval cases.

An Abstraction of Low Level Video Features for Automatic Retrievals of Explosion Scenes

This paper proposes an abstraction mechanism of the low-level digital video features for the automatic retrievals of the explosion scenes from the digital video library. In the proposed abstraction mechanism, the regional dominant colors of the key frame and the motion energy of the shot are defined as the primary low-level visual features of the shot for the explosion scene retrievals. The regional dominant colors of shot are selected by dividing its key frame image into several regions and extracting their regional dominant colors, and the motion energy of the shot is defined as the edge image differences between key frame and its neighboring frame. Upon the extensive experimental results, we could argue that the recall and precision of the proposed abstraction and detecting algorithm are about 0.8, and also found that they are not sensitive to the thresholds.