In [1]:

    

from numpy import *
from PIL import *
import pickle
from pylab import *
import os


    

In [2]:

    

import sift
import dsift
dsift = reload(dsift)
import imtools
imtools = reload(imtools)


    

In [3]:

    

def read_gesture_features_labels(path):
    # make a list of the files with .dsift at the end
    featlist = [os.path.join(path, f) for f in os.listdir(path)
               if f.endswith('.dsift')]
    
    # read features
    features = []
    for featfile in featlist:
        l, d = sift.read_features_from_file(featfile)
        features.append(d.flatten())
    features = array(features)
    
    # generate labels
    labels = [featfile.split('/')[-1][0] for featfile in featlist]
    
    return features, array(labels)


    

In [4]:

    

features, labels = read_gesture_features_labels('train/')
test_features, test_labels = read_gesture_features_labels('test/')
classnames = unique(labels)


    

In [5]:

    

# the first letter of the file name is the label
print labels


    

    




['F' 'C' 'A' 'V' 'A' 'B' 'A' 'F' 'V' 'P' 'F' 'P' 'V' 'C' 'A' 'V' 'V' 'C'
 'B' 'V' 'C' 'C' 'A' 'F' 'V' 'P' 'V' 'P' 'C' 'A' 'F' 'A' 'C' 'B' 'P' 'B'
 'B' 'F' 'V' 'B' 'A' 'C' 'B' 'C' 'C' 'V' 'A' 'B' 'A' 'P' 'P' 'P' 'F' 'B'
 'B' 'A' 'C' 'A' 'B' 'F' 'P' 'C' 'A' 'A' 'V' 'A' 'B' 'P' 'F' 'P' 'A' 'B'
 'V' 'F' 'B' 'A' 'F' 'F' 'C' 'V' 'B' 'V' 'C' 'F' 'P' 'P' 'A' 'F' 'P' 'P'
 'F' 'F' 'F' 'V' 'V' 'A' 'C' 'C' 'F' 'P' 'F' 'A' 'V' 'F' 'C' 'B' 'V' 'P'
 'B' 'C' 'P' 'V' 'P' 'F' 'V' 'V' 'C' 'A' 'B' 'C' 'F' 'P' 'C' 'B' 'V' 'P'
 'C' 'V' 'P' 'C' 'C' 'B' 'A' 'A' 'C' 'P' 'C' 'P' 'B' 'F' 'F' 'B' 'P' 'A'
 'A' 'C' 'F' 'P' 'V' 'C' 'B' 'V' 'B' 'F' 'B' 'A' 'V' 'C' 'F' 'B' 'F' 'V'
 'B' 'V' 'A' 'P' 'A' 'V' 'P' 'F' 'B' 'P' 'A' 'B']

In [6]:

    

import knn


    

In [7]:

    

def print_confusion(res, test_labels, classnames):
    
    n = len(classnames)
    
    class_ind = dict([(classnames[i], i) for i in range(n)])
    
    confuse = zeros((n, n))
    for i in range(len(test_labels)):
        confuse[class_ind[res[i]], class_ind[test_labels[i]]] += 1
    
    print 'Confusion matrix for'
    print classnames
    print confuse


    

In [9]:

    

# k neighbors
accs = []
for k in arange(1, 10):
    knn_classifier = knn.KnnClassifier(labels, features)

    res = array([knn_classifier.classify(test_features[i], k) for i in
            range(len(test_labels))])

    # accuracy
    acc = sum(1.0*(res==test_labels))/len(test_labels)
    print 'Accuracy:', acc
    accs.append(acc)


    

    




Accuracy: 0.716346153846
Accuracy: 0.605769230769
Accuracy: 0.519230769231
Accuracy: 0.475961538462
Accuracy: 0.427884615385
Accuracy: 0.365384615385
Accuracy: 0.331730769231
Accuracy: 0.288461538462
Accuracy: 0.269230769231
0.269230769231

In [10]:

    

print accs


    

    




[0.71634615384615385, 0.60576923076923073, 0.51923076923076927, 0.47596153846153844, 0.42788461538461536, 0.36538461538461536, 0.33173076923076922, 0.28846153846153844, 0.26923076923076922]

In [22]:

    




    

In [ ]: