---->Your Transformation---->
In this homework you will implement convolution and local processing. Your program will read images from file. Then you will apply convolution with some matrix and save the result in a file. Local algorithms will also be implemented in similar manner but they will require operations other than addition and multiplication.
Your program should run in two modes:
Your program will be tested as follows:
parta in.ppm out.ppm b <---- this invocation applies a blur mask to image parta in.ppm out.ppm -f mask.txt <---- this invocation applies the mask in 'mask.txt' |
mask.txt format is same as ppm format BUT entries are floating point numbers NOT integers. so the following is valid:
3 3 -0.1 -0.1 -0.1 -0.1 0.8 -0.1 -0.1 -0.1 -0.1 |
You must take care of overflow/underflows in the resulting image. You can check your results with common image processing programs like Paint Shop/Gimp.
In this part you will enhance the images using the methods you've learnt. The image given contains a hidden message. You have to enhance it to make the message readable. Your program should be able to enhance any such images, having similar distortions.
| source image (in.ppm) |
---->Your Transformation----> |
original image (out.ppm)
???????? |
Your programs should be run as:
partb in.ppm out.ppm <--- here 'in.ppm' is the source image and 'out.ppm' is the image you generated |
In second set of images you have to combine low resolution color images and high resolution gray level image to obtain a better color image. You can use any of the methods you've learnt. Also feel free to experiment with different methods. Write details in your report.
Hi-res grayscale image (ingray.ppm) |
---> your transformation ---> |
Desired output image (out.ppm) |
Lo-res color image (incolor.ppm) |
Your programs should be run as
partc incolor.ppm ingray.ppm out.ppm <--- here 'incolor.ppm' is the lowres color image, 'ingray.ppm' is the hires grayscale image and 'out.ppm' is the image you generated |
For part A you use following files:
Convert these images to ppm format using your favorite image manipulation program.
For parts B&C you use following files:
or all zipped in one file
Here you can find an implementation of canny edge detector:
tar.gz all files and use submit command to submit all files. Include a Makefile to build your sources and a readme file for additional explanations.
Write a report about the results you obtain. Discuss methods and results. Explain methods you use. Bring report to my room (A206) before deadline.
For questions first use newsgroup.