http:^^www.cs.cornell.edu^info^faculty^brian_smith.html
来自「This data set contains WWW-pages collect」· HTML 代码 · 共 272 行 · 第 1/2 页
HTML
272 行
MIME-Version: 1.0
Server: CERN/3.0
Date: Wednesday, 20-Nov-96 18:57:08 GMT
Content-Type: text/html
Content-Length: 10260
Last-Modified: Friday, 31-May-96 14:15:26 GMT
<html><head><title> Faculty Research Interests : Brian Smith </title></head><body><H1> Brian Smith </H1><!WA0><!WA0><!WA0><!WA0><!WA0><!WA0><!WA0><!WA0><!WA0><!WA0><!WA0><!WA0><!WA0><!WA0><!WA0><!WA0><IMG SRC="http://www.cs.cornell.edu/Info/Faculty/bsmith/bsmith-small.gif"><I><DL><DT> bsmith@cs.cornell.edu<DT> Xerox Professor of Computer Science<DT> Ph.D., University Of California at Berkeley, 1994<DT> Office: 4107B Upson Hall<DT> Office phone: 607-255-1180<DT> Office hours this semester: Tues & Thurs 3:00 to 4:00</DL></I><ul><li><!WA1><!WA1><!WA1><!WA1><!WA1><!WA1><!WA1><!WA1><!WA1><!WA1><!WA1><!WA1><!WA1><!WA1><!WA1><!WA1><A HREF="#Research">Research Interests</A><li><!WA2><!WA2><!WA2><!WA2><!WA2><!WA2><!WA2><!WA2><!WA2><!WA2><!WA2><!WA2><!WA2><!WA2><!WA2><!WA2><A HREF="#Teaching">Teaching</A><li><!WA3><!WA3><!WA3><!WA3><!WA3><!WA3><!WA3><!WA3><!WA3><!WA3><!WA3><!WA3><!WA3><!WA3><!WA3><!WA3><A HREF="#Publications">Selected Publications</A><li><!WA4><!WA4><!WA4><!WA4><!WA4><!WA4><!WA4><!WA4><!WA4><!WA4><!WA4><!WA4><!WA4><!WA4><!WA4><!WA4><A HREF="#Talks">Research Talks</A><li><!WA5><!WA5><!WA5><!WA5><!WA5><!WA5><!WA5><!WA5><!WA5><!WA5><!WA5><!WA5><!WA5><!WA5><!WA5><!WA5><A HREF="#Links">Misc Links</A></ul><A NAME = "Research"><H3>Research Interests</H3>My research goal is to make video a first class data type in ourcomputing environment. To this end, my research group,<!WA6><!WA6><!WA6><!WA6><!WA6><!WA6><!WA6><!WA6><!WA6><!WA6><!WA6><!WA6><!WA6><!WA6><!WA6><!WA6><A HREF="http://www.cs.cornell.edu/Info/Projects/multimedia/zeno.html"> Project Zeno</A>,is building technologies supporting the storage, communication, andprocessing of continuous media data. In contrast to other commercialand research approaches, which require specialized hardware, operatingsystems, or networks to be usable, all of these technologies we aredesigning fit into the current research environment. Our premise isthat the current hardware, software, and communication infrastructureis sufficient to support research into continuous media systems andapplications. We are verifying this hypothesis by building workingsystems.<p>Our research on storage systems is directed towards building the Zenodistributed video file server. The Zeno architecture uses a network ofworkstations connected by a generic local area network (e.g., anethernet), a common environment in computing research laboratories.Each workstation can act simultaneously as both a client and a serverof continuous media data. As a client, a workstation plays videostored on one or more servers. As a server, a workstation is a fileserver for video data. Each client can receive video stored at severalservers, and each server can service several clients. Compared withlarge centralized servers, the advantages of this design are:<ol><li> Scalability. As new clients are added, new servers are automatically added.<li> Load balancing. The load generated by serving videos is distributed both across machines and across networks (in the case where the servers are located on different networks).<li> Low initial investment. By utilizing existing infrastructure, the Zeno architecture promotes early adoption in research environments with almost no initial investment.</ol><p>Our research on communication systems is centered around best effortdelivery protocols. Such protocols are built on existing networkprotocols and, in contrast to many other research efforts, do not needto reserve network resources to establish a connection. Resourcereservation protocols are well suited to the national communicationinfrastructure where users can be charged on a per call basis forbandwidth and connections, but are poorly suited to networkenvironments where the network is a shared resource equally accessibleby all researchers. Our approach is appropriate for the latterenvironments, commonly found in research laboratories. Thecommunication protocol we have developed, called Cyclic-UDP, is builton top of the UDP datagram protocol, and is designed to transport audioand video data in playback applications in local, metropolitan, andwide area networks. Cyclic-UDP is used by the Zeno file server todeliver audio and video data to clients. A<!WA7><!WA7><!WA7><!WA7><!WA7><!WA7><!WA7><!WA7><!WA7><!WA7><!WA7><!WA7><!WA7><!WA7><!WA7><!WA7><A HREF="http://www.cs.cornell.edu/Info/Faculty/bsmith/nossdav.ps.gz"> paper describing cyclic UDP is availableonline </A>, as well as the<!WA8><!WA8><!WA8><!WA8><!WA8><!WA8><!WA8><!WA8><!WA8><!WA8><!WA8><!WA8><!WA8><!WA8><!WA8><!WA8><A HREF="http://www.cs.cornell.edu/Info/Faculty/bsmith/Talks/CUDPTalk.html"> slides from a research talk. </A><p>Our research on processing video data has been two-fold. First, we aredeveloping algorithms to process video data in the compressedrepresentation. Processing video without decompression leads todramatic speed-ups in processing performance since it both removes thetime-consuming processes of compression and decompression and reducesthe amount of data that must be processed. Experiments with animplementation of these ideas on JPEG compressed image data indicatesthat the data can be processed one to two orders of magnitude fasterthan what was possible with previous approaches. We are currentlyextending these ideas by parallelizing the algorithms using networks ofworkstations and by developing a method for transcoding video insoftware. In video transcoding, video is translated from onecompression format to another, a useful operation for video fileservers that must service heterogeneous clients.A <!WA9><!WA9><!WA9><!WA9><!WA9><!WA9><!WA9><!WA9><!WA9><!WA9><!WA9><!WA9><!WA9><!WA9><!WA9><!WA9><A HREF="http://www.cs.cornell.edu/Info/Faculty/bsmith/Global-Effects.ps.Z"> paper describing compresseddomain processing </A> is available online.<p>Our research on video processing is also exploring methods to simplifyexperimentation with video processing by developing a programminglanguage where video is a first class data type. This language, calledRivl (pronounced "rival"), allowsvideo processing effects to be specified independent of the resolutionand format of the source material. The language does for video whatPostscript did for text and graphics: it provides a resolutionindependent method for specifying video processing. Thus, the sameprogram can process low quality QuickTime video very quickly whileediting decisions are made, and then be used to format a high qualityfinished product off-line, in much the same way that Postscript can bepreviewed on a workstation at low quality, then sent to a 2600 dpiprinter for camera ready copy. A <!WA10><!WA10><!WA10><!WA10><!WA10><!WA10><!WA10><!WA10><!WA10><!WA10><!WA10><!WA10><!WA10><!WA10><!WA10><!WA10><A HREF="http://www.cs.cornell.edu/Info/Faculty/bsmith/rvl-tcl.ps"> paper describing RVL </A>is available online.<p>A <!WA11><!WA11><!WA11><!WA11><!WA11><!WA11><!WA11><!WA11><!WA11><!WA11><!WA11><!WA11><!WA11><!WA11><!WA11><!WA11><A HREF="http://www.cs.cornell.edu/Info/Faculty/bsmith/Talks/UMinn.html"> talk that reviews our research onvideo processing, </A> both compressed domain processing and Rivl,is available online.<p>⌨️ 快捷键说明
复制代码Ctrl + C
搜索代码Ctrl + F
全屏模式F11
增大字号Ctrl + =
减小字号Ctrl + -
显示快捷键?