why k means do not rectangle the red area drawn by harris corner?

[Ho Yeung Lee]

i would like to ask whether 

at line
dd = dst>0.01*dst.max()

and
the part of code below from script, whether write wrong in array access or assignment

    testing1 = zip(*np.where(dd == 1))
    locations = testing1
    testing1 = [list((float(elem[0]),float(elem[1]))) for elem in testing1]
    centroids,_ = kmeans(testing1,10)
    row,col = vq(testing1,centroids)
    locationsgroup = {}
    keyone = len(locationsgroup)+1
    for jj in range(0, len(row)):
        for ii in range(0, len(np.nonzero(row == jj)[0])):
            locationsgroup = addtogroupkey(locationsgroup, keyone, locations[np.nonzero(row == jj)[0][ii]])
        keyone = len(locationsgroup)+1
        #for ii in range(0, len(np.nonzero(row == jj)[0])):
            #locationsgroup = addtogroupkey(locationsgroup, keyone, locations[np.nonzero(row == jj)[0][ii]])
        #keyone = len(locationsgroup)+1
    colortolocation[eachcolor] = locationsgroup





On Tuesday, August 15, 2017 at 4:33:51 AM UTC+8, jlad... at itu.edu wrote:
> On Monday, August 14, 2017 at 12:30:21 PM UTC-7, Ho Yeung Lee wrote:
> > https://gist.github.com/hoyeunglee/df7e6cb9b76c576b26fd2bb2b26bfe2f
> > 
> > sample image
> > https://drive.google.com/file/d/0Bxs_ao6uuBDUY09qM1JMWS1Ob0k/view?usp=sharing
> > 
> > would like to rectangle bound the eye  ,mouth corner etc
> > which detected by harris corner
> > 
> > but using kmeans can not rectangle the red area drawn by harris corner
> 
> Does your code run without errors?  You appear to be asking more general question, about machine learning and image processing algorithms, than about Python.