EE 639 Advanced Topics in Signal Processing and Communication

Fall 2004: Multimedia Information System

[Home] [Syllabus] [Lectures] [Readings] [Homework] [Project Recommendations] [Final Projects] [Announcements]

Final projects from Fall 2004

A Wiener filter approach for removal of ringing artifacts from JPEG coded images

Nathir A. Rawashdeh and Il-Won Shin

JPEG is a widely used image compression standard. It allows the encoding of images at less bits per pixel than in the original image reducing their file sizes up to ten times. This benefit comes at the price of reduced image quality, and at low bit rates, the visual delectability of blocking and ringing compression artifacts. In this report for EE639 of Fall'2004, we review the research literature regarding blocking and ringing artifacts reduction for both the BDCT-based JPEG and the newer wavelet-based JPEG 2000 formats. In addition, we present initial results of a Wiener filtering approach for the reduction of ringing artifacts in JPEG images. We discuss how this approach can be extended to not require the original, artifact-free, image and how it could be applied to imaging applications.

Any-Time, Any-Place Television

Jesus Caban

Television has an important influence in our lives. This being true, people are interested in the idea of a multimedia system that allows them to watch television at any time and in any place. An international group, started by the TV-AnyTime Consortium in 1999, is trying to produce standards for storage-based TV receivers such as Personal Video Recorders. Its goal is to standardize the infrastructure for Digital Video Recording (DVR) devices like TiVo. As mobile devices become more and more popular, there is a growing demand to watch television in portable devices like cell phone or PDA. In this project we will build a system that can receive and record live broadcast television, process the digitized video, and stream the recorded video to computers and display devices such as personal computers, laptops and iPAQ hand-held.

Hiding Privacy Information in Video Surveillance System

Wei Zhang

This paper proposes a detailed framework of storing privacy information in surveillance video as a watermark. Authorized personnel is not only removed from the surveillance video as in [1] but also embedded into the video itself, which can only be retrieved with a secrete key. A perceptual-model-based compressed domain video watermarking scheme is proposed to deal with the huge payload problem in the proposed surveillance system. A signature is also embedded into the header of the video as in [2] for authentication. Simulation results have shown that the proposed algorithm can embed all the privacy information into the video without affecting its visual quality. As a result, the proposed video surveillance system can monitor the unauthorized persons in a restricted environment, protect the privacy of the authorized persons but, at the same time, allow the privacy information to be revealed in a secure and reliable way.

A Survey of Experiential Computing

Kungyen Chang

The proliferation of large amounts of heterogeneous data for specific events or experiences demand a new type of system to organize and present data in a form that is most useful for the user. Experiential computing addresses this need, and is examined in the form of an experiential meeting system, surveillance and monitoring systems, and a multimedia management system. Limitations of this approach, as well as future possibilities are discussed.

A Survey of Face Recognition Technologies

Dharshan Charith Medonza

For this project I did a survey of the technologies for Human Face Recognition. Face recognition is an important skill that we acquire. It plays a major role in conveying identity and emotion. The human ability to recognize faces is remarkable. We can recognize thousands of faces learned throughout our lifetime. But what about computational methods for face recognition? How can we get a computer to recognize a human face? For this I will go through some computational methods that have been used for human face recognition.

Video Clip Detection Using Sequence Matching Technique

Veera Ganesh and Vijay Venkatesh

The objective of this project is to develop an automatic sequence matching technique based on histogram and ordinal feature to detect the presence of a video clip in a given video database. The feature extraction is done in real time on broadcast television. A simple KD-tree index structure is built based on the extracted features to robustly search short clips of a given duration in large video collection. We implement a K-NN Query and Range query to efficiently search the database. Experiments show that the implemented algorithm performs efficiently in detecting the presence of the video clip in a given database.

Sen-ching Samson Cheung
Last modified: Tue Aug 24 11:20:57 EDT 2004