I'm developing a image editor by webgl. I did some research with the existing apps which contains image dehaze effect, FotoJet and Polarr. Here are the example images exported by FotoJet and Polarr.
Original:
FotoJet:
Polarr:
So, what's the dehaze algorithm they used? I think their algorithm is different, can anyone help me figure out the algorithms?
I find the algorithm of Kaiming He at http://kaiminghe.com/cvpr09/. I think it's the solution.
The FotoJet and Polarr don't implement the perfect result as Kaiming He's Demo. I'll try it by myself.
Related
First of all, thanks for reading my question. I'm beginner in computer vision.
I read a lot but I didn't find any solution.
I have an image and I want to detect logo/logos on it.
Also, I have a whole of images with different logos, all image containing a logo on it and nothing more.
Can you help me with any idea of how to detect logo/logos on an image when I have a whole (thousands) of training sets (known logos set)?
It can be done by using the SURF or SIFT feature detection algorithm for few known logos, by matching the given image with all of the others but I have a huge dataset, and I can't match with all other images.
To try all images in the dataset takes toooooo much time :)
Can be useful any SDK? (it can be even for mobile phones or for desktop also).
Or can I use some multiple algorithms for it?
I found an interesting paper about this question with a SIGMA algorithm, but I can't find any description for these algorithms (http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=5495345).
I think to detect the features on the images is OK (SIFT, maybe SURF).
But I think the problem is with the big number of known images/logos.
I think it should be stored in a special way.
Ex. made a tree somehow from the thousand of known logos, or to separate them in groups.
Is it possible to do this task?
I appreciate any help.
The thousands of training sets is useful only to test your algorithm, it will not help to analyze a new image.
I made a bit of pattern recognition in the past, I would start this way: look for sharp edges (sharp color transitions too). So an edge filter and statistical analysis about features all located in the same corner. The result of the algorithm will be a number that you will use with your training set.
Since you are doing original reserch be prepared for a long work. If a SDK with a function "ImageHasLogo()" exists yet, you will find it on Google.
Hey so I was wondering if anyone has any reading suggestions for papers/algorithms on how to render scenes in the style that was used for the video game Okami.
I've done some searching on my own and have found some "sumie" style papers but the algorithms presented there are black and white and I'm looking for something with color.
Any help is appreciated!
You're looking for Non-photorealistic rendering. There are dozens of algorithms, I can't tell you which one is used in the game exactly, it might be more algorithms together. Anyway, you may find some algorithms and techniques on Stylized Depiction in Computer Graphics and by searching on google..
I am curious if there are libraries already developed for creating these noises in javascript?
I have searched the web for about two hours looking for libraries in javascripts. I have seen this implementation for the simplex noise:
https://gist.github.com/304522
but I have not seen any libraries made up specifically for javascript.
If no one knows of any libraries then link suggestions on how to develop these functions would be great as well. I am currently doing Google searches for everything I have asked here I just figured better to have multiple web crawlers helping me out concurrently :) Also if someone on stack overflow has already found a quality site then even better.
Thanks guys!
Here you can find a thoroughly explanation of the algorithm, also the original code from Worley (explained in Texturing and modelling. A procedural approach) ported to Processing.
At the moment I'm working deploying to Javascript.
For the tiling part of your question: if you have a Perlin or Simplex noise function you don't really need a library to create tiling images because the math is really not that difficult. What is difficult is understanding how it works. For instance: a spherical noise map is way easier than simple 2D tiling image because you need to work with four dimensions for the latter.
I once wrote a blog post explaining in detail how to create tiling noise (including Javascript code examples).
It shows you how to tile in linear, cylindrical, planar and spherically (including cubical panorama).
I've made a javascript app that renders perlin/simplex noise to an html5 canvas, you can find it here: http://lencinhaus.github.com/canvas-noise
Well it's not a proper library, but the javascript code contains implementations of value noise, perlin noise (both in classic and improved versions) and simplex noise, all derived from the original Perlin's C implementations (or a more optimized java version for simplex noise, see the references in the code). You can tweak the noise parameters visually and immediately see the results in the browser, so maybe it can be useful for you.
hope that helps!
Recently I've been messing about with algorithms on images, partly for fun and partly to keep my programming skills sharp.
I've just implemented a 'nearest-neighbour' algorithm that picks n random pixels in an image, and then converts the colour of each other pixel in the image to the colour of its nearest neighbour in the set of n chosen pixels. The result is a kind of "frosted glass" effect on the image, for a reasonably large value of n (if n is too small then the image gets blocky).
I'm just wondering if anyone has any other good/fun algorithms on images that might be interesting to implement?
Tom
This book, Digital Image Processing, is one of the most commonly used books in image processing classes, and it will teach you a lot of basic techniques that will help you understand other algorithms better, like the ones Ants Aasma suggested.
Try making an Andy Warhol print. It's pretty easy in Java. For more ideas, just look at the filters available in GIMP or a similar program.
Marching Squares is a computer vision algorithm. Try using that to convert black and white raster images to object based scenes.
Turns the image into a pizza
Take N images, relate them via an MC-Escher-style painting
"Explode" an image from the inside out
Convert the image into a single-color blocks (piet-style) based on all the colours within.
How about tie-dye algorithm?
Fun to toy with and easy to code filters are:
kaleidoscope
lens
twirl
There are a lot of other filters, but especially the kaleidoscope gives much bang for the bucks. I have made my own graphics editor with lots of filters and is also looking for inspiration.
Instead of coding image filters, I personally would love to code Diffusion Curves, but unfortunately have little time for fun.
If you want to try something more challenging look for SIGGRAPH papers on the web. There are some really nifty image algorithms presented at that conference. Seam carving is one cool example that is reasonably straightforward to implement.
If you want something more challenging try to complete the symmetry of broken objects
Recently I saw demonstration of image distortion algorithm that preserves some objects in image. For example changing weight/height ratio but preserving people faces in image to look realistically. The problem is i can't find the reference to that demonstration neither I know the name of such transformations.
Can anybody point me to a references to such algorithms?
Perhaps you are referring to liquid rescale?
Seam carving is the current favourite. Liquid Rescale is an implementation of it.