http:^^www.cs.cornell.edu^info^projects^csrvl^csrvlmeng.html
来自「This data set contains WWW-pages collect」· HTML 代码 · 共 108 行
HTML
108 行
MIME-Version: 1.0
Server: CERN/3.0
Date: Monday, 25-Nov-96 00:16:39 GMT
Content-Type: text/html
Content-Length: 4952
Last-Modified: Tuesday, 19-Sep-95 22:26:57 GMT
<TITLE>Suggested Robotlab M.Eng Projects </TITLE> <HEAD><H1>Suggested Robotlab M.Eng Projects </H1></HEAD> <li> Robotics<ol><li> <em>Wireless <!WA0><!WA0><!WA0><!WA0><a href="http://www.cs.cornell.edu/Info/Projects/csrvl/mobot.html">Robotics</a></em>-Get <!WA1><!WA1><!WA1><!WA1><a href="http://www.cs.cornell.edu/Info/Projects/csrvl/pics/rose.gif">Rosemary</a> to communicate with the Sun workstations using the wireless radio modem, with wireless video (using a cheap box whichsends an analog video signal over radio (like TV)). Hook this up tothe ATM video source and build an application of realtime visiontracking/etc. with rosemary as the host. (Matt)<li> <em>Visual <!WA2><!WA2><!WA2><!WA2><a href="http://www.cs.cornell.edu/Info/Projects/csrvl/mobot.html">Robot</a> Navigation</em>-An extension to Tom's project:Mapmaking with visual cues. Can use ideas from (1), but the robotwalks around, and, using depth and position information produces a mapof the world, uses it to navigate. (Matt) [Tom's project: a set of navigational becons were created for the robots]</ol><li> Video Processing and Machine Vision<ol><li> <em>Automatic Storyboarding</em> - use the <!WA3><!WA3><!WA3><!WA3><a href="http://www.cs.cornell.edu/Info/People/rdz/dissolve.html">dissolve detector</a> code toautomate generation of movie storyboards, possible outputing htmlpages that show the story board. (Justin and Ramin)<li> <em>Parallelization of the Mpeg encoder</em> Write a 'real-time'(or at least very fast) Mpeg encoder using the ATM cluster/SP2 and ATMvideo source. Jon Swartz and Brian Smith were interested in this. (Matt)<li> <em>Port Split-C to Windows NT</em>. This could prove to benon-trivial, as many aspects of the SMP design rely on a UNIX processenvironment, in particular ELF, and shared memory. Use of NT featuressuch as threads won't work. (Matt and Justin V.)<li> <em>Parallel RIVL</em>. Providing Split-C/Active Messages supportfor RIVL. We'd like RIVL to be the general framework for going imagecomputation, display, applications building, etc. (It should helpproductivity a lot if we have an easy way for people to experiment andvisualize vision algorithms, instead of writing crufty C programs.)This project would provide fine-grain data parallel processing supportfor the RIVL architecture, basically by (a) porting various RIVLimage-processing primitives to Split-C (for automatic parallelization)and (b) allowing users to write new RIVL primitives in Split-C forlinking into the system. The easiest way to do all of this, I think:Port RIVL itself to Split-C (should be easy) and make existing RIVLbehaviour a subset of the "Split-RIVL" (that is, local processing onone machine). (Matt)<li> <em> Effects of Compression</em> What (empirical) effects do various compression algorithms have on common vision algorithms. (Justin and Ramin)<li> <em> Edge Clustering</em> Develop algorithm for clustering ofedges that represent a single object. (Justin and Ramin)<li> <em> Adobe Photoshop plug-ins</em>(Nawaaz)</ol> <li> Algorithms <ol><li> <em>Optimization</em>Create an environmentfor the automatic optimization of highly iterative imagetransformation algorithms using the m4 macro language and perlscripts. Write m4 macros to automatically produce C-code according tosome set of parameters (like how many times to unroll a loop) and thewrite a perl script to compile the set of programs and determine whichis the fastest by timing the transformation on some set of images.This would greatly speed the production of optimized code for a givenmachine. I have examples started of this idea for the rank transform.<li><em> Search-based compilation</em> Modelling compilation as a search problem on a system with constraints :<p>The system will consist of<ul> <li> a set of processors <li> a set of memories <li> an interconnect network <li> a set of constraints<ul> <li> number of processors <li> size of memory <li> which processor can access which memories in a cycle <li> ordering imposed by the operations to be performed. <li> etc.</ul></ul>The code to be parellised will be tight loops and hence small. The idea is to search the space of complied code to find the optimal one.<p>This is most probably NP complete, but speed is not the essence here.(Nawaaz)<li><em> Performance Simulations</em> The simulation is with respectto the algorithms we are interested in. The many aspects that can besimulated are : a) Cache -- implement and study cache accesses onvarious algorithms. Here we assume a single processor. b) SharedMemory Multiprocesssors -- simulate and study the effect of busbandwidth, memory architecture , cache coherence on the system (withrepect to alogrithms like correlation etc) (Nawaaz)<li><em> DSPs</em> Implementation / Compilation / Code generation forDSP chips (MPV?) : This is more specific then <em> Search-basedcompilation</em>. Just to get an idea of what the chips are reallycapable of.</ol> </ul>⌨️ 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?