Unable to produce the results as per the paper

[yash1996]

Hi YAksoy

I was unable o produce the same results as given in the paper on any other image and also new to MATLAB.

• I used the SIGGRAPH18SSS repo to generate the features that gave me docia.mat file.
• I later placed the mat file along with the input PNG in this repo and ran the demo.m where I read the docia.png file as the input and the results that I got was as follows.

Which is fine but was unable to understand how to get the middle section of the PNG that is highlighted as follows.

Also I am not sure if the code was even using the docia.mat file. And it just used the PNG file below.

Could you help me on how to run this to get the input file from the features(docia.mat file) as desired.

Regards
Yash

[joelwebb]

Yash, I am having a difficult time running any of the Matlab files for the demo. Did you have to change the original code to run it correctly? Hopefully if i could get it going I might be able to help you find the intermediate image.

[lph529372693]

Hi YAksoy

I was unable o produce the same results as given in the paper on any other image and also new to MATLAB.

• I used the SIGGRAPH18SSS repo to generate the features that gave me docia.mat file.
• I later placed the mat file along with the input PNG in this repo and ran the demo.m where I read the docia.png file as the input and the results that I got was as follows.

Which is fine but was unable to understand how to get the middle section of the PNG that is highlighted as follows.

Also I am not sure if the code was even using the docia.mat file. And it just used the PNG file below.

Could you help me on how to run this to get the input file from the features(docia.mat file) as desired.

Regards
Yash

I can help you solve the mat file problem. Check call the preprocessFeatures. m function, in the command line window input:
tmp = load('docia.mat');
features = tmp.embedmap;
image= imread('docia_one.png');%The original figure
simp = preprocessFeatures(features, image)
imwrite(simp,'docia_feature.png','png')%So this is the three-dimensional feature image that you want

docia.png in the source code is a patchwork of the original image and the feature image.
In the command line window input:
image = imread('docia_one.png');
features = imread('docia_feature.png');
com_image = [image, features];
imshow(com_image)
imwrite(com_image,'docia.png');

But I haven't been able to produce the segmentation effect like you, my effect is very bad.How are your parameters set?Can you share your matlab code to me?

[Lakshaysharma048]

@yash1996 how did you run demo.m without error. i am seeing people getting stuck at the error" unable to find "adjacentRegionsGraph".

how did you overcome this issue?