Introductory Computer Organization
In this lab assignment, you will implement two image processing kernels in MIPS assembly. The input
for both functions is a square sized grayscale image. Refer to the handout(补充内容在附件中） on
memory layout to learn how a 2D array is stored in memory. In this case, images are stored in
a 2D array where each element is a byte that represent each pixel’s intensity (0x00 used for bla
to 0xFF. Output image in both parts will have the same size as the input image. We have included
a utility function that load and store image data from images stored on disk with pgm file format (a
simple raw format). Your task is to write the body of the functions that loads an array of pixels
along with its dimensions and process them and stores the result back in memory.
Fig. 1. Image processing pipeline
1
https://docs.google.com/document/d/1odXDX-j4CVW_tkrIpqHtgQz9ArhDeMiKyMSKSs0Pn0s/edit?usp=sharing
Part 1) Image Thresholding:
Image thresholding is a simple way of partitioning an image into a foreground and background. This
image analysis technique is a type of image segmentation that isolates objects by converting
grayscale images into binary images. An example of such conversion is shown below:
Lena.png Lena_binary.png
Your function should replace the value of each pixel in the grayscale image with the maximum
possible value (0xFF, i.e. brightest value), if the intensity (value) of that pixel is greater than or equal
to some constant T (function input), or otherwise with the minimum value (0x00, i.e. darkest value).
Remember that these values are unsigned.
Part 2) Affine Transformation:
For the second part, you are implementing a function to apply an affine transformation to the input
image. Such transformation is described using a 2×3 transform matrix (M) that map pixels at input
image to different coordinates in output image. Affine transformations have this property that they
map parallel lines in input image to parallel lines in output image. Examples of affine transformations
include: translation, rotation, reflection and scale.
The following algorithm demonstrates applying affine transformation to an input image. In this
algorithm x and y represents current pixel coordinates — x being the column index and y being the
row index. Top-left pixel has coordinates of (0, 0).
or each input image pixel at x, coordinatef y :
M M y Mx0 = oo × x + o1 × + 02
M M y My0 = 1o × x + 11 × + 12
utput(x, ) input(x0, y0) if cols and rows y otherwise 0o y = > xo ≥ 0 > 0 ≥ 0
Although image pixel are bytes, matrix elements are integers. Here’s some examples of affine
transformations:
2
https://en.wikipedia.org/wiki/Transformation_matrix#Affine_transformations
Sample input
1 0 0
0 1 0
0 0 1
Identity matrix
2 0 0
0 1 0
0 0 1
Scale matrix (x = 2)
0 1 0
1 0 0
0 0 1
Rotation matrix (+ translate)
3
1 1 0
0 1 0
0 0 1
Shear matrix
Testing
Lab3.s comes with a tester which tests your function with proper parameters such as input arrays and
images. If your program passes tests on arrays correctly, you are guaranteed to receive full points for this
lab.
To test your code against images, you need to modify the source code to point to input images location on
your computer. “Lenna.pgm” is the input image and “lenna_thresh.pgm”, “lenna_scaled.pgm”,
“lenna_shear.pgm” and “lenna_scale.pgm” are output images that will be generated by parts 1 and 2 of this
lab. You need to verify them manually.
fin: .asciiz “PATH/TO/lenna.pgm”
fout_thresh: .asciiz “PATH/TO/lenna_thresh.pgm”
fout_rotate: .asciiz “PATH/TO/lenna_rotation.pgm”
fout_shear: .asciiz “PATH/TO/lenna_shear.pgm”
fout_scale: .asciiz “PATH/TO/lenna_scale.pgm”
4