region_search.py

import numpy as np
import os 
#np.set_printoptions(threshold=np.NaN)


def find_biggest_region(data):
	

	region_inf=np.zeros(shape=(31,31,31))
	region_flag=[0]
	for ii in range(1,30):
		for jj in range(1,30):
			for zz in range(1,30):
				if data[ii,jj,zz]==1:
					neighbor=np.array([region_inf[ii-1,ll,mm] for ll in range(jj-1,jj+2) for mm in range(zz-1,zz+2)]+[region_inf[ii,jj-1,zz-1],region_inf[ii,jj,zz-1],region_inf[ii,jj-1,zz]])
					valid_neighbor=neighbor[neighbor!=0]
					if len(valid_neighbor)==0:
						region_inf[ii,jj,zz]=max(region_flag)+1
						region_flag.append(max(region_flag)+1)
					if len(valid_neighbor)!=0:
						new_flag=min(valid_neighbor)
						region_inf[ii,jj,zz]=new_flag
						for flag in set(valid_neighbor):
							region_inf[region_inf==flag]=new_flag
							region_flag.remove(flag)
						region_flag.append(new_flag)

	num=0
	region_flag.remove(0)
	
	if len(region_flag)==0:
		return data,0

	biggest_region=region_flag[0]
	for flag in region_flag:
		temp_num=np.sum(region_inf==flag)
		if temp_num>num:
			num=temp_num
			biggest_region=flag

	out=np.zeros(shape=(31,31,31),dtype=int)
	out[region_inf==biggest_region]=1
	return out,len(region_flag)