Scribd
Upload
68 views2 pages

Handwriting Recognition With Python

The document discusses handwriting recognition using scikit-learn. It loads handwritten digit data and analyzes sample images and pixels. Various machine learning classifiers are applied including random forest. Sample and validation data is used to score the random forest classifier at around 89% accuracy.

Uploaded by

ravikantssharma
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
68 views2 pages

Handwriting Recognition With Python

The document discusses handwriting recognition using scikit-learn. It loads handwritten digit data and analyzes sample images and pixels. Various machine learning classifiers are applied including random forest. Sample and validation data is used to score the random forest classifier at around 89% accuracy.

Uploaded by

ravikantssharma
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

12/7/2016 PyRevolution/Handwriting_Recognition.

ipynb at master · snazrul1/PyRevolution · GitHub

Handwriting Recognition
Author: Syed (Sadat) Nazrul

Import libraries and dataset from Scikit­Learn

In [13]: import [Link] as plt 
%matplotlib inline 
from [Link] import load_digits 
 
digits = load_digits() 

Analyze a sample image

In [14]: import pylab as pl  
[Link]()  
[Link]([Link][0])  
[Link]()  

<[Link] at 0x7f1b8c186550>

Analyze image pixels

Each element represents the pixel of our greyscale image. The value ranges from 0 to 255 for an 8 bit pixel.

In [15]: [Link][0] 

Out[15]: array([[  0.,   0.,   5.,  13.,   9.,   1.,   0.,   0.], 
       [  0.,   0.,  13.,  15.,  10.,  15.,   5.,   0.], 
       [  0.,   3.,  15.,   2.,   0.,  11.,   8.,   0.], 
       [  0.,   4.,  12.,   0.,   0.,   8.,   8.,   0.], 
       [  0.,   5.,   8.,   0.,   0.,   9.,   8.,   0.], 
       [  0.,   4.,  11.,   0.,   1.,  12.,   7.,   0.], 
       [  0.,   2.,  14.,   5.,  10.,  12.,   0.,   0.], 
       [  0.,   0.,   6.,  13.,  10.,   0.,   0.,   0.]])

Visualize first 15 images

In [16]: images_and_labels = list(zip([Link], [Link])) 
[Link](figsize=(5,5)) 
for index, (image, label) in enumerate(images_and_labels[:15]): 
    [Link](3, 5, index + 1) 
    [Link]('off') 
    [Link](image, cmap=[Link].gray_r, interpolation='nearest') 
    [Link]('%i' % label) 

[Link] 2/4
12/7/2016 PyRevolution/Handwriting_Recognition.ipynb at master · snazrul1/PyRevolution · GitHub

In [17]: import random 
from sklearn import ensemble 
 
#Define variables 
n_samples = len([Link]) 
x = [Link]((n_samples, ‐1)) 
y = [Link] 
 
#Create random indices  
sample_index=[Link](range(len(x)),len(x)/5) #20‐80 
valid_index=[i for i in range(len(x)) if i not in sample_index] 
 
#Sample and validation images 
sample_images=[x[i] for i in sample_index] 
valid_images=[x[i] for i in valid_index]
 
#Sample and validation targets 
sample_target=[y[i] for i in sample_index] 
valid_target=[y[i] for i in valid_index]
 
#Using the Random Tree Classifier 
classifier = [Link]() 
 
#Fit model with sample data 
[Link](sample_images, sample_target) 
 
#Attempt to predict validation data 
score=[Link](valid_images, valid_target) 
print 'Random Tree Classifier:\n'  
print 'Score\t'+str(score) 

Random Tree Classifier: 
 
Score  0.889429763561 

In [20]: i=150 
 
[Link]()  
[Link]([Link][i])  
[Link]()  
[Link](x[i]) 

<[Link] at 0x7f1b804b0cd0>

Out[20]: array([0])

© 2016 GitHub, Inc. Terms Privacy Security Status Help Contact GitHub API Training Shop Blog About

[Link] 3/4

You might also like