How do i map mu euclidean distances calulcated in order to create stack to make image foveated

Question

SANJIT SINGH 2022-12-12

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1876142-how-do-i-map-mu-euclidean-distances-calulcated-in-order-to-create-stack-to-make-image-foveated

回答： Karan Singh 2023-9-5

%image files

i1 = imread("iron-man-marvel-chrome-theme-wallpaper.png");

i2 = imread("iron-man-marvel-chrome-theme-wallpaper2.png");

i3 = imread("iron-man-marvel-chrome-theme-wallpaper3.png");

i4 = imread("iron-man-marvel-chrome-theme-wallpaper4.png");

i5 = imread("iron-man-marvel-chrome-theme-wallpaper5.png");

%concatened array full of images

m = cat(4,i1,i2,i3,i4,i5);

%check each rgb invidually

red = m(:,:,1);

green = m(:,:,2);

blue = m(:,:,3);

%create/update rgb stacks, rgb stacks will be 5

red = m(:,:,1,:);

green = m(:,:,2,:);

blue = m(:,:,3,:);

r = squeeze(red);

g = squeeze(green);

b = squeeze(blue);

red = m(:,:,1,1);

x1 = 306;

y1 = 404;

x = 1:353;

y = 1:483;

%Returns 2-D grid coordinates based on the coordinates contained in vectors x and y.

%X is a matrix where each row is a copy of x, & Y is a matrix where each column is a copy of y

[X,Y] = meshgrid(y,x);

distance = sqrt((X-x1).^2 + (Y-y1).^2);

I am trying to map my distance variable to image with highest resolution and basically create fixation points of when i focus on anothe rpoint within image based on distances from center origin, I want to apply all distance calulcated from center origin and make image fovaeated. output of the distance variable stores pixels and want to map those pixels to my image to make image foveated

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Karan Singh 2023-9-5

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1876142-how-do-i-map-mu-euclidean-distances-calulcated-in-order-to-create-stack-to-make-image-foveated#answer_1302031

在 MATLAB Online 中打开

Hi Sanjit,

From what I understand, the code you provided loads five image files (“i1”, “i2”, “i3”, “i4”, and “i5”) into an array called “m”. Each image is represented as a 3D matrix, where the third dimension represents the RGB channels.

Next, you extracted the red, green, and blue channels from the “m” array and stored them separately in “red”, “green”, and “blue” variables. Then, you created/updated the RGB stacks by assigning the extracted channels back to “red”, “green”, and “blue”.

After that, you assigned the first image's red channel (“m(:,:,1,1)”) to the “red” variable, and defined the coordinates (“x1” and “y1”) for the center origin.

To create fixation points based on distances from the center origin, you calculated the Euclidean distance from the center origin to each pixel in the image using “meshgrid”. The resulting distance values are stored in the “distance” variable.

To map the distance variable to an image and create a foveated effect, you can use the distance values as indices to select pixels from the image. 

Here's an example of how you can achieve this:

% Load the highest resolution image 
highest_resolution_image = i5; 
% Define the center origin coordinates 
x1 = 306; 
y1 = 404; 
% Create a meshgrid for the image coordinates 
[x, y] = size(highest_resolution_image(:,:,1)); 
[X, Y] = meshgrid(1:y, 1:x); 
% Calculate the distance from the center origin 
distance = sqrt((X - x1).^2 + (Y - y1).^2); 
% Normalize the distance values between 0 and 1 
normalized_distance = distance / max(distance(:)); 
% Apply the foveated effect by selecting pixels based on distance 
foveated_image = highest_resolution_image; 
for channel = 1:3 
    foveated_image(:,:,channel) = uint8(double(foveated_image(:,:,channel)) .* normalized_distance); 
end 
% Display the foveated image 
imshow(foveated_image); 

In this code, we load the highest resolution image (“i5” in this case) and define its center origin coordinates (“x1” and “y1”). We create a “meshgrid” for the image coordinates using the size of the image. Then, we calculate the distance from the center origin for each pixel using the “meshgrid”.

Next, we normalize the distance values between 0 and 1 to ensure they are within the valid range for pixel intensity. We apply the foveated effect by multiplying the pixel values of each channel by the normalized distance.

Finally, we display the resulting foveated image using “imshow”.

Attached below are some documentation links that you may find helpful: