| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Help Conquer Cancer
Thread: Help Conquer Cancer Paper in News & Update |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 7
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
|
||
|
|
nasher
Veteran Cruncher USA Joined: Dec 2, 2005 Post Count: 1423 Status: Offline Project Badges: 
|
Wow that was an interesting article thanks for posting it. alot of technical information on it but it is nice to know what we are doing here is actually getting useful results in real world application
---------------------------------------- |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi.
Great news, any advance in fighting cancer is a good thing. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Many thanks , great news !!
----------------------------------------[Edit 1 times, last edit by Former Member at May 11, 2010 12:37:32 AM] |
||
|
|
rilian
Veteran Cruncher Ukraine - we rule! Joined: Jun 17, 2007 Post Count: 1460 Status: Offline Project Badges: 
|
it is not clear to me, whether this 80% success detection rate done after current 45% of crunching (so we need 55% of crunching for further 20%) or not?
---------------------------------------- |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
rilina, talking about this bit?
----------------------------------------Technical Abstract: We have developed an image-analysis and classification system for automatically scoring images from high-throughput protein crystallization trials. Image analysis for this system is performed by the Help Conquer Cancer (HCC) project on the World Community Grid. HCC calculates 12,375 distinct image features on microbatch-under-oil images from the Hauptman-Woodward Medical Research Instituteâs High-Throughput Screening Laboratory. Using HCC-computed image features and a massive training set of 165,351 hand-scored images, we have trained multiple Random Forest classifiers that accurately recognize multiple crystallization outcomes, including crystals, clear drops, precipitate, and others. The system successfully recognizes 80% of crystal-bearing images, 89% of precipitate images, and 98% of clear drops. Reads to me as: Out of 100 tasks we do of the crystal-bearing type gives an 80 are a hit Out of 100 tasks that are filtered of precipitate images type gives 89 returns Out of 100 of the clear drops types are properly identified, 98 are marked correctly. (think there was a clear drop discussion some time ago) Don't know the mix, but if even, it works out as 89% success-rate for the project.
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 1 times, last edit by Sekerob at May 11, 2010 2:48:49 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi rilian,
----------------------------------------HCC is an ongoing project. This is just a report of where the scientists were back when they wrote the paper. I take for granted that they are hard at work evolving the software to give better results. Lawrence Added: I will stick my neck out a bit and pass on some unsubstantiated rumor. I could be completely incorrect. The HCC people have been looking at coding for GPUs on this project. As ATI has been releasing its own GPU language, they have been writing and rewriting, learning how to do it. But the learning curve has been arduous, so the current thinking is to finish up the WCG project to refine the algorithm, then look into transferring it to GPU for speed. This is just my own interpretation of cryptic comments. We remain committed to GPU computing, but we are not currently under time pressure to deploy a GPU project, so we will concentrate our upgrades to accommodate oncoming projects. [Edit 1 times, last edit by Former Member at May 11, 2010 9:50:45 PM] |
||
|
|
|