The folder raw_data contains minimal image examples and their human full interpretation data (referred below as annotations). The raw annotations for each minimal image category are stored in a MAT file. Each annotated image is represented as a MATLAB struct element with two fields: 'mirc', which is the MIRC image and 'human_interpretation', which contains annotations for the MIRC image.
The human_interpretation field has three cells. The first is a list (represented as cells) of annotated [y,x] points. The second cell is the list of contours and the third is a list of regions. Each contour is a matrix of size n by 2, where n is the number of sampled [y,x] points in the contour. Each region is stored as a vector of [top row, bottom row, left column, right column].
The zip file contains also a short MATLAB script for plotting interpretations.
Extract raw_data.zip and update the file CONSTS.py to your current folder locations.
To visualize a minimal image and its full interpretation run, e.g.,
$python mat2segmap.py -n HORSE_HEAD -i 13
where -n is input arg for mirc object name, and -i is the mirc image index.
To create the full dataset of mirc images and their interpretations (as segmentation maps)
update the file CONSTS.py to your current folder locations, and then run:
$python gen_mirc_interp_dataset.py
This should create a ~12M size folder containing mirc images and their segmentation maps.