Chinese Students Develop PhotoSketch - A Real Magic Wand
I read that Google was a project of some graduate students. Now here comes another group in the same clan, they may not become another Larry page s(founder of Google) but they created buzz and ripples on the web world to get enough of attention. So here's the story --
I am one of the worst artists, in my childhood(in teenage - adult hood also) I used to get the worst marks in the drawing class (ofcourse I used to score very well in academics ). Then I used to think , if I had some magic pencil - by which if I draw some rough lines it will create a beautiful painting for me. Ofcourse I never got such type of magic pencil , but here comes a magic software which do the same thing.
The program, which has been built by five Chinese students at Tsinghua University and the National University of Singapore, uses a vast library of images to turn a basic sketch in to a photograph. A website that allows people to experiment with the software has crashed under the volume of traffic.
Few hours before posting this message I was able to see their website , but now it crashed.. so much ripples it created on the net. It also appeared on the top in Google Hot trends (thats how I got to know of it ).
Many computer websites, techo-bloggers described the stuff as just "mind blowing" and some commenters started questioning "from which era these guys arrived?"
To keep in simple pictures what it mean....
So if you draw the match stick drawing with some text labels like the one on the left , you will get the picture like the one on the right . Amazing, mind blowing , magic of the decade , what not . So this is the software what the chinese students have created.
The authors of the tool – Tao Chen, Ming-Ming Cheng, Ping Tan, Ariel Shamir, and Shi-Min Hu – say that the software can take any rough sketch showing the shape of each object, labelled with a name, and find images that correspondent to each element. It can also judge between images to decide which is the best match for the picture, and then stitch all of the elements together in a single image, adding shadows to give a naturalistic look and feel.
The abstract they posted reads like this
"We present a system that composes a realistic picture from a simple freehand sketch annotated with text labels. The composed picture is generated by seamlessly stitching several photographs in agreement with the sketch and text labels; these are found by searching the Internet. Although online image search generates many inappropriate results, our system is able to automatically select suitable photographs to generate a high quality composition, using a filtering scheme to exclude undesirable images. We also provide a novel image blending algorithm to allow seamless image composition. Each blending result is given a numeric score, allowing us to find an optimal combination of discovered images. Experimental results show the method is very successful; we also evaluate our system using the results from two user studies."
The video that shows how it works is here - (may be this also be down with traffic)
It is really tough to separate a person or thing from it's back ground by using wand, cutting tools on editing softwares like photoshop, fireworks, corel.. they made it automatic and too blending all images on one page.. huff...
Web users have been quick to point out that the tool is initially most likely to be used to create funny images, perhaps showing celebrities in amusing or compromising positions. But technology experts have been quick to recognise the potential of the software.
So before making my own view and comment , I want to make it clear that I am still not out of the SHOCK and still ammused that wheter this can be really done... let those wonder guys put the stuff again somewhere so that I can have hands on it...
if you are a programming geek and if you want to have a look ,they also provided source code.. click here (this link is also presently down given the huge traffic on the website ) ,( I will not any how touch it.. not that it may be in chinese.. but frankly I don't have intelligence to understand how the output works.. then dare not think what the code is )
Now people like me can become Da VIncis