Auto cell detection by RICM, for 20150206 FC1 fixed

auto_detect.py

# auto_detect.py

from ij import IJ
from ij.plugin import ImageCalculator

IJ.run("Close All")
IJ.run("Stack From List...", "open=/Users/hiroyuki/Desktop/images.txt")
IJ.run("Gaussian Blur...", "sigma=5 stack")
imp = IJ.getImage()
imp.setTitle("all")

def statistics():
    # How many points are in our dataset?
    num_points = 10
    
    # For each of those points how many dimensions do they have?
    dimensions = 2
    
    # Bounds for the values of those points in each dimension
    lower = 0
    upper = 200
    
    # The K in k-means. How many clusters do we assume exist?
    num_clusters = 3
    
    # When do we say the optimization has 'converged' and stop updating clusters
    opt_cutoff = 0.5
    
    # Generate some points
    points = [kmeans.makeRandomPoint(dimensions, lower, upper) for i in xrange(num_points)]
    
    # Cluster those data!
    clusters = kmeans.kmeans(points, num_clusters, opt_cutoff)

    # Print our clusters
    for i,c in enumerate(clusters):
        for p in c.points:
            print " Cluster: ", i, "\t Point :", p


def detect(original,fr,to,count):
	IJ.run(original,"Slice Keeper", "first=%d last=%d increment=1"%(fr,to))
	imp = IJ.getImage()
	name = 'kept%d'%count
	imp.setTitle(name)
	IJ.run("Invert", "stack")
	IJ.run("Z Project...", "projection=[Average Intensity]")
	ave = IJ.getImage()
	IJ.selectWindow(name)
	res = ImageCalculator().run("Subtract create stack", imp, ave)
	# IJ.selectWindow("Result of kept1")
	IJ.run(res,"Auto Threshold", "method=Otsu white stack")
	IJ.run(res,"Invert", "stack")
	IJ.run(res,"Set Measurements...", "centroid stack redirect=None decimal=3")
	IJ.run(res,"Analyze Particles...", "size=100-Infinity display stack")
	res.show()

detect(imp,1,21,1)
detect(imp,22,39,2)
detect(imp,40,62,3)
detect(imp,63,87,4)