Graduation design to do image segmentation and recognition of something.
Look at the paper and see Superpixels start the brain complement is the pixel interpolation calculated
Then the more I look at it, the more I find it, the other thing.
http://blog.sina.com.cn/s/blog_50363a7901011dtd.html
have defined
Many existing algorithms in computer vision use the Pixel-grid as the underlying representation. For example, stochastic models of images, such as Markov the random fields, is often defined on the this regular grid. Or, face detection is typically do by matching stored templates to every fixed-size (say, 50x50) windows in the image.
The Pixel-grid, however, is a natural representation of visual scenes. It is rather a "artifact" of a digital imaging process. It would is more natural, and presumably more efficient, to work with perceptually meaningful entities obtained from a low -level grouping process. For example, we can apply the normalized Cuts algorithm to partition a image into, say, segments Ixels).
The result of the low-level feature grouping process is called superpixels .
Empirica l studies and applications] "title=" mega-pixel: empirical based research and application [superpixel: empirical studies and applications] "Action-data=" http%3a%2f%2fttic.uchicago.edu% 2f~xren%2fresearch%2fsuperpixel%2fimages%2f100048.jpg "action-type=" Show-slide "style=" margin:0px; padding:0px; border-width:0px; List-style:none; " |
Empirical studies and applications] "title=" mega-pixel: empirical based research and application [superpixel: Empirical studies and applications] "Action-data=" http%3a%2f%2fttic.uchicago.edu%2f~ Xren%2fresearch%2fsuperpixel%2fimages%2f100048_seg.jpg "action-type=" Show-slide "style=" margin:0px; padding:0px; border-width:0px; List-style:none; " > |
(a) |
(b) |
Empirica l studies and applications] "title=" mega-pixel: empirical based research and application [superpixel: empirical studies and applications] "Action-data=" http%3a%2f%2fttic.uchicago.edu% 2f~xren%2fresearch%2fsuperpixel%2fimages%2f100048_overseg.jpg "action-type=" Show-slide "style=" margin:0px; padding:0px; border-width:0px; List-style:none; " |
Empirical studies and applications] "title=" mega-pixel: empirical based research and application [superpixel: Empirical studies and applications] "Action-data=" http%3a%2f%2fttic.uchicago.edu%2f~ Xren%2fresearch%2fsuperpixel%2fimages%2f100048_seg_overseg.jpg "action-type=" Show-slide "style=" margin:0px; padding:0px; border-width:0px; List-style:none; " > |
(c) |
(d) |
A is the image to be processed, B is the human Visual Division, C is Superpixel Map,d is based on C new Human visual segmentation
There are many ways to
Introduction to the http://blog.sina.com.cn/s/blog_6f126f5b010134a3.html method
What is a hyper-pixel superpixels thing?