<?xml version="1.0" encoding="utf-8"?>
<feed xml:lang="en" xmlns="http://www.w3.org/2005/Atom"><title>Recent posts to news</title><link href="https://sourceforge.net/p/c-a-i-r/news/" rel="alternate"/><link href="https://sourceforge.net/p/c-a-i-r/news/feed.atom" rel="self"/><id>https://sourceforge.net/p/c-a-i-r/news/</id><updated>2008-10-17T05:01:16Z</updated><subtitle>Recent posts to news</subtitle><entry><title>Speed Demon</title><link href="https://sourceforge.net/p/c-a-i-r/news/2008/10/speed-demon/" rel="alternate"/><published>2008-10-17T05:01:16Z</published><updated>2008-10-17T05:01:16Z</updated><author><name>Brain_ReCall</name><uri>https://sourceforge.net/u/userid-1902296/</uri></author><id>https://sourceforge.netcc36989e55763d6660f099fbbc0388760230e96c</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;Well, well, well. CAIR v2.16 managed to pull a rabbit out of the hat. I reworked my threading so that things remain persistent instead of destroying them after each seam. On smaller images, I found a speed increase of just over 50%. The larger the images, the less the poor threading hurt it, but its still above 20%. If I would have known that was such a serious issue it would have been dealt with long ago. Oh well.&lt;/p&gt;
&lt;p&gt;On a related note, I'm probably going to be shelving CAIR for awhile. I'm going to try to start CAVR - Content Aware Video Retargeting. CAVR will hopefully implement a few ideas from the doctors and bring seam carving to videos. I already have quite a few ideas for it. Perhaps the biggest challenge is finding a suitable video library that I can use for testing purposes (much in the same way I just use EasyBMP for CAIR).&lt;br /&gt;
My requirements for a video library would be:&lt;br /&gt;
- Open source C++&lt;br /&gt;
- Cross-platform&lt;br /&gt;
- The ability to grab a pixel value from a frame&lt;br /&gt;
- Read and write SOME kind of video format (I really don't care which, would prefer uncompressed AVI)&lt;br /&gt;
- Easy to use&lt;/p&gt;
&lt;p&gt;If you have any suggestions, please email me at brain.recall@gmail.com&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>Not dead yet</title><link href="https://sourceforge.net/p/c-a-i-r/news/2008/05/not-dead-yet/" rel="alternate"/><published>2008-05-09T21:36:25Z</published><updated>2008-05-09T21:36:25Z</updated><author><name>Brain_ReCall</name><uri>https://sourceforge.net/u/userid-1902296/</uri></author><id>https://sourceforge.netdfb172b4ead60e76555647dd8318576e81a2633b</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;I'm still around and still poking at CAIR. Speed increases will be coming, as well as a few changes, and possibly a few new features.&lt;/p&gt;
&lt;p&gt;Hopeful wishlist:&lt;br /&gt;
* Put all internal maps into one, reducing memmove() overhead. This gave about a 5% speed boost.&lt;br /&gt;
* Make the number of threads a complier option (including energy map threads).&lt;br /&gt;
* Make the threads persistent to reduce OS overhead in creating and destroying threads over and over.&lt;br /&gt;
* Add a new averaging technique to also allow pixels above and below the pixels on the seem to be utilized.&lt;br /&gt;
* Make separate source and destination weight matrices parameters, rather than using the input as also the output as it is now.&lt;/p&gt;
&lt;p&gt;That's the short list for now. I'll probably try to wrap up the performance improvements into a new release, and work on the threading issues in a later one. &lt;/p&gt;
&lt;p&gt;Please email me with ideas and suggestions at brain.recall@gmail.com&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>New developer and CVS</title><link href="https://sourceforge.net/p/c-a-i-r/news/2007/12/new-developer-and-cvs/" rel="alternate"/><published>2007-12-01T15:45:57Z</published><updated>2007-12-01T15:45:57Z</updated><author><name>Brain_ReCall</name><uri>https://sourceforge.net/u/userid-1902296/</uri></author><id>https://sourceforge.net1d971d8e731dda709528f9ef3dc0e07bed1195da</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;I would first like to welcome Jib to the developers staff. He provided a lot of the changes into v2.11. He also has an idea on how to optimize Energy_Map(), so those changes will be going into the next release.&lt;/p&gt;
&lt;p&gt;The code is now available under CVS. This will make it far easier for me and Jib to collaborate on these upcoming changes, and it allows everyone to grab the latest-n-greatest. &lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>Fixed it</title><link href="https://sourceforge.net/p/c-a-i-r/news/2007/11/fixed-it/" rel="alternate"/><published>2007-11-15T01:23:27Z</published><updated>2007-11-15T01:23:27Z</updated><author><name>Brain_ReCall</name><uri>https://sourceforge.net/u/userid-1902296/</uri></author><id>https://sourceforge.neted002cbd14ea8f6df961bdd80c7f0e26e0f0bb45</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;I had a major slowdown going to v2.7 while enlarging. The problem was a function called Add_Weights(), which I didn't yet optimize. The function previously ran best with column-major layout, and switching over to row-major sent the cache misses through the roof. Fixed it, and did a little multi-threading to it for a minor boost.&lt;/p&gt;
&lt;p&gt;I'm stuck at implementing Poisson image reconstruction, the last feature in the paper that I haven't implemented. If you're a math nut, and are willing to sacrifice some brain power, read this: &lt;a href="http://portal.acm.org/citation.cfm?id=882269" rel="nofollow"&gt;http://portal.acm.org/citation.cfm?id=882269&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Please email me, or post here on the boards, if you know how to apply this to the seam-carving method.&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>Something happened</title><link href="https://sourceforge.net/p/c-a-i-r/news/2007/11/something-happened/" rel="alternate"/><published>2007-11-06T00:11:59Z</published><updated>2007-11-06T00:11:59Z</updated><author><name>Brain_ReCall</name><uri>https://sourceforge.net/u/userid-1902296/</uri></author><id>https://sourceforge.net00514ef3f505bbe096ffd32c4b066df1f91c4823</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;The new version 2.7 I touted as about 25% faster than v2.6.2. Apparently, that isn't the case.&lt;br /&gt;
I have been performing all of my speed tests with the same image performing the same action. Just today I noticed that I lost some performance during enlarging, which for the most part I never pay attention to since I assumed performance boosts seen for removal will be seen for adding as  well.&lt;/p&gt;
&lt;p&gt;Right now I noticed a 20% speed reduction from v2.6.2 to v2.7. How going row-major has caused this I'm uncertain at this point. I noticed on my dual-core machine that one core seems to be stalling at about 50% CPU usage with v2.7, but under v2.6.2 they both are at full. &lt;/p&gt;
&lt;p&gt;I'm going to look into this, but I was hoping someone else could confirm this on a different CPU architecture (I'm currently a Athlon 64 X2). For some reason I believe this might be a cache issue, but I'm not sure just yet.&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>Compiling Tips and a Plea For Help</title><link href="https://sourceforge.net/p/c-a-i-r/news/2007/11/compiling-tips-and-a-plea-for-help/" rel="alternate"/><published>2007-11-03T23:17:53Z</published><updated>2007-11-03T23:17:53Z</updated><author><name>Brain_ReCall</name><uri>https://sourceforge.net/u/userid-1902296/</uri></author><id>https://sourceforge.net5ea2cb407d4810f84ffe8bb08cac3457209b37da</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;I did a little more performance analyzing, and on the tip from a friend I went and tried the Intel C++ compiler. Sure enough, after kicking on all of the fancy optimization switches to their max, I managed to get an additional speed improvement. Resizing with my standard test scenario (1024x678 image down to 800x600) I found about a 10% improvement with quality == 0 (from the 11-10 second range down to the 10-9 second range), and about a 20% improvement with quality == 1 (15 seconds down to the 12 second range). &lt;/p&gt;
&lt;p&gt;I obviously recommend using this compiler when available. Unfortunately, it’s only freely available for the Linux platform, and since I don’t have the funds for a license I probably won’t be releasing a compiled DLL for the Windows platform. This was more of an exercise in discovery than anything.&lt;/p&gt;
&lt;p&gt;I have a new version in the works, but no major changes (there’s a left-over Temp object in CAIR_Add() that I forgot to get rid of). Honestly, I’m really running out of ideas for performance improvements. The only thing I could think of was maybe trying to add at least one additional thread to Energy_Map(). Also, I want to try to figure out what Intel did to my convolve loops to get another 10% out of them.&lt;/p&gt;
&lt;p&gt;Again, this is another plea to anyone who’s a good C++ hacker to find some possible improvements to my code (speed-wise at least). &lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>Halloween Surprise </title><link href="https://sourceforge.net/p/c-a-i-r/news/2007/11/halloween-surprise-/" rel="alternate"/><published>2007-11-01T03:21:59Z</published><updated>2007-11-01T03:21:59Z</updated><author><name>Brain_ReCall</name><uri>https://sourceforge.net/u/userid-1902296/</uri></author><id>https://sourceforge.netd4634e2afc58c97b29a72e08804f701407e2729b</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;The new v2.6 release gives about a 15% speed boost. This, I believe, solidly locks it to one of the fastest seam carvers. Please let me know if this isn't the case. Resizing a 1024x678 image down to 800x600 now takes 11 seconds with quality == 0, and just about 15 seconds with quality == 1. CML now actually resizes logically.&lt;/p&gt;
&lt;p&gt;Unfortunately, this release might mark a milestone. I'm running out of areas to optimize. As of right now, Energy_Map() consumes about 70% of the processing time. I will obviously devote most of my time in that area, but I'm not sure on how to squeeze out more performance as of right now. &lt;br /&gt;
My latest tirade involved using some SSE instructions, and although they work, I didn't receive any noticeable advantage. I was using a packed minimum operation, to perform four min_of_three() operations simultaneously. But, that operation works on floats, so another SSE instruction was need to convert the integers to floats and back again. SSE4 has an integer minimum, but that’s so new no one has a processor that supports it yet. Also, unaligned memory movement (or, conversely, no standardized byte-boundary declarations between compilers) pretty much shoots it down. I may look into it again, but I don’t expect the vast improvements I was hoping for.&lt;/p&gt;
&lt;p&gt;If looking through my code you have any idea for an area of improvement, have no hesitation to email me or post in the SourceForge forums for this project. If you have an idea on how to modify the algorithm for some improvement, also, let me know.&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>Latest and Greatest</title><link href="https://sourceforge.net/p/c-a-i-r/news/2007/10/latest-and-greatest/" rel="alternate"/><published>2007-10-27T21:31:02Z</published><updated>2007-10-27T21:31:02Z</updated><author><name>Brain_ReCall</name><uri>https://sourceforge.net/u/userid-1902296/</uri></author><id>https://sourceforge.net391b3415f66049e354120b735fad864ed5a23fa7</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;A bunch of changes went through version 2.5. The biggest change was reworking how the CML worked internally, which gave speed improvements throughout. CAIR is now one of the fastest implementations that I can find, and with very low quality (such as just 5%, or 0.05), its still extremely fast but produces great results. It's almost becoming practical.&lt;/p&gt;
&lt;p&gt;I also added a new feature, CAIR_Removal(), which attempts to automatically remove areas marked for removal and then return the image to its original size. From some of my simple tests it seems to work pretty well, but let me know how it goes for you.&lt;/p&gt;
&lt;p&gt;Also sometime here I want to move most of the information on the Google Page's website into the Wiki for Sourceforge. The Google-Pages website is also in need of an update, so I might of well move it while I'm at it.&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>Seam Carving GUI</title><link href="https://sourceforge.net/p/c-a-i-r/news/2007/10/seam-carving-gui/" rel="alternate"/><published>2007-10-25T00:07:56Z</published><updated>2007-10-25T00:07:56Z</updated><author><name>Brain_ReCall</name><uri>https://sourceforge.net/u/userid-1902296/</uri></author><id>https://sourceforge.net5b3165c39df818699fdde1994afaacd24181ad5e</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;Gabe Rudy informed me that he has got a working GUI that now uses CAIR. See: &lt;a href="http://gabeiscoding.com/" rel="nofollow"&gt;http://gabeiscoding.com/&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;All that it needs is a faster CAIR backend, which the next release will be.&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>Version 2.0</title><link href="https://sourceforge.net/p/c-a-i-r/news/2007/10/version-20/" rel="alternate"/><published>2007-10-21T18:24:49Z</published><updated>2007-10-21T18:24:49Z</updated><author><name>Brain_ReCall</name><uri>https://sourceforge.net/u/userid-1902296/</uri></author><id>https://sourceforge.net930889d8290eaa2552563cafc5ffd8af58087415</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;CAIR has undergone a substantial remake. Internally, it no longer depends on the EasyBMP library. In fact, I only use EasyBMP to read and write files in my sample program. I'm now using an custom made template class for all internal matrices. This means, though, that a translator function will probably have to be written from whatever image system your using to the internal one CAIR uses.&lt;/p&gt;
&lt;p&gt;BUT, CAIR has gained a 50% performance boost across the board. With quality == 1 and resizing a 1024x678 image down to 800x600 now takes about 41 seconds compared to the old 81 seconds. With quality == 0, thats down to 21 seconds from the old 38.&lt;/p&gt;&lt;/div&gt;</summary></entry></feed>