Website:
https://www.eecis.udel.edu/~compbio/FigSplit
Our paper at:
https://doi.org/10.1093/bioinformatics/btx611
Main function:
Command: matlab /code/FigSplit.m
Inputs:
Image_file_path: The path to the image that needs separation.
Outputs:
The identified panels (in jpg format) from the compound image and their coordinates (in text format) at the original image.
Explainations of submodules called by FigSplit.m
Submodule 1: Image preprocessing (Lines 8-31 in FigSplit.m)
Inputs:
Im_original: The original image that needs separation.
Outputs:
Im0: The preprocessed image.
Submodule 2: Connected component analysis (/code/white_margin_detection.m)
Inputs:
Im0: The preprocessed image.
Outputs:
pointsforseparation: The coordinates of identified panels by using the connected component analysis method.
Submodule 3: Separation quality assessment (/code/evaluation.m)
Inputs:
Im0: The preprocessed image.
pointsforseparation: The coordinates of identified panels.
Outputs:
The coordinates of identified panels after the evaluation and self-correction process.
Submodule 4: Handle stitched image (/code/get_white_seg_recursion_function.m)
Inputs:
Im0: The preprocessed image.
Outputs:
The coordinates of identified panels by choosing the peak position of the horizontal projection or vertical projection.
Submodule 5: Handle blurry and fragmented image (/code/CCAuseSuan1.m)
Inputs:
Susan_im: The image generated by applying SUSAN edge detector to the preprocessed image im0.
Outputs:
The coordinates of identified panels by applying CCA method to Susan_im.
\subsection{Datasets}
The datasets used for the Figure Separation task at ImageCLEF2013, ImageCLEF2015, ImageCLEF2016.
Shield:
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
