view labelimage2points.py @ 0:39e6d6a84257 draft

planemo upload for repository https://github.com/BMCV/galaxy-image-analysis/tools/labelimage2points/ commit c3f4b766f03770f094fda6bda0a5882c0ebd4581
author imgteam
date Sat, 09 Feb 2019 14:36:17 -0500
parents
children 07525a7d9ea0
line wrap: on
line source

import argparse
import sys
import pandas as pd
import skimage.io
from skimage.measure import label
from skimage.data import checkerboard
import numpy as np



def labelimage2points(input_file):
    img_in = skimage.io.imread(input_file) 
        
    #amount of regions
    amount_label = np.max(img_in)
    
    # iterate over all regions in order to calc center of mass
    center_mass = []
    for i in range(1,amount_label+1):    
        #get coordinates of region
        coord = np.where(img_in==i)
        # be carefull with x,y coordinates
        center_mass.append([np.mean(coord[1]),np.mean(coord[0])])

    #make data frame of detections
    out_dataFrame = pd.DataFrame(center_mass)


    #return
    return(out_dataFrame)


if __name__ == "__main__":
    parser = argparse.ArgumentParser()
    parser.add_argument('input_file', help='input file')
    parser.add_argument('out_file', help='out file (CSV)')

    args        = parser.parse_args()
    input_file  = args.input_file
    out_file    = args.out_file

    #TOOL
    out_dataFrame = labelimage2points(input_file)

    #Print to csv file
    out_dataFrame.to_csv(out_file, index=False, header=False, sep="\t")