This folder contains the dataset used in the paper:

Directing User Attention via Visual Flow on Web Designs
Xufang Pang*, Ying Cao*, Rynson W.H. Lau and Antoni B. Chan (*Joint first authors)
ACM Trans. on Graphics (Proc. ACM SIGGRAPH Asia 2016), Dec. 2016

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
The dataset include a set of webpages with following information:

1. Image of each webpage
2. Segmentation of page components on each webpage
3. Eye gaze data of multiple viewers under task-driven condition on each webpage
4. Category of each webpage

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Data structure of the dataset is described as below:

Webpages are stored in a cell array (webpages).

Each cell is a webpage with following fields:
		
	.img: Image of the webpage.

	.components: Segmentation of page components on the webpage.
		         A cell array where each cell is a page component with following field:
                    .polygon: coordinates of four corner points of the page compnent
                        .x: x-coordinates 
                        .y: y-coordinates
	
	.eyeXYD: Eye gaze data of multiple viewers on the webpage. 
		     A cell array where each cell is eye gaze data of a viewer stored in a N-by-3 array:
                1. each row is a gaze point [X Y D] (X: x-coordinate, Y: y-coordinate, D: duration in millisecond)
                2. order of a gaze point in the array indicates its temporal order during eye tracking process

	.category: Category of the webpage (email, fileshare, job, product, shopping, social)		
