Handwritten Digit Recognition Based on Modified LLE Algorithm

2014 ◽  
Vol 602-605 ◽  
pp. 2290-2293
Author(s):  
Bo Yu Gu ◽  
Ye Chun Li
Keyword(s):  
Euclidean Distance ◽  
Locally Linear Embedding ◽  
Learning Approach ◽  
Affine Transformations ◽  
Accuracy Rate ◽  
Handwritten Digit Recognition ◽  
Digit Recognition ◽  
Handwritten Digit ◽  
Linear Embedding ◽  
Locally Linear

Locally linear embedding is an efficient manifold learning approach. A modified locally linear embedding algorithm is proposed to cope with the interferences of affine transformations in handwritten digit recognition. In order to offset all kinds of affine transformations, the Euclidean distance is replaced by the tangent distance which is more appropriate for handwritten digit recognition based on image. And the number of neighborhood is computed automatically based on the similarity of images. Experimental results show that the accuracy rate of is improved.

scholarly journals Research on Mnist Handwritten Numbers Recognition based on CNN

2021 ◽  
Vol 2138 (1) ◽  
pp. 012002
Author(s):  
Yang Gong ◽  
Pan Zhang
Keyword(s):  
Training Set ◽  
Accuracy Rate ◽  
Handwritten Digit Recognition ◽  
Data Set ◽  
Digit Recognition ◽  
Input Layer ◽  
Handwritten Digit ◽  
Model Training ◽  
Increasing Demand ◽  
Fully Connected

Abstract In view of the increasing demand for handwritten digit recognition, a handwritten digit recognition model based on convolutional neural network is proposed. The model includes 1 input layer and 2 convolutional layers (5*5 convolution Core), 2 pooling layers (2*2 pooling core), 1 fully connected layer, 1 output layer, and use the mnist data set for model training and prediction. After a lot of training and participation, the accuracy rate of the training set was finally reached to 100%, and the accuracy rate of 99.25% was also achieved on the test set, which can meet the requirements of recognizing handwritten digits.

An Improved Locally Linear Embedding Method for Feature Extraction

2011 ◽  
Vol 467-469 ◽  
pp. 487-492
Author(s):  
Wei Zhang ◽  
Wei Jia Zhou
Keyword(s):  
Feature Extraction ◽  
Tangent Space ◽  
Euclidean Distance ◽  
Classical Method ◽  
Locally Linear Embedding ◽  
Linear Embedding ◽  
Simulation Results ◽  
Space Distance ◽  
Better Than ◽  
Locally Linear

In this work, a feature extraction approach based on improved Locally Linear Embedding(LLE) is proposed. In the algorithm, tangent space distance is introduced to LLE, which overcomes the shortcoming of original LLE method based on Euclidean distance. It can satisfy the requirement of locally linear much better and so can express the I/O mapping quality better than classical method. Simulation results are given to demonstrate the effectiveness of the improved LLE method.

Application of Locally Linear Embedding Based on Improved Distance in Neuron Classification

2014 ◽  
Vol 926-930 ◽  
pp. 2996-2999
Author(s):  
Zhen Zhen Wang ◽  
Xiao Jun Tong ◽  
Shan Zeng
Keyword(s):  
Euclidean Distance ◽  
Locally Linear Embedding ◽  
Support Vector ◽  
Morphological Parameters ◽  
Svm Algorithm ◽  
Good Classification ◽  
Linear Embedding ◽  
Sample Component ◽  
Locally Linear

For locally linear embedding (LLE) algorithm of the shortcoming, an improved distance algorithm LLE is proposed, in locally linear embedding algorithm the distribution of sample component is different and the Euclidean distance can’t reflect sample distance actually. In the experiment, a sample of 231 neurons is obtained, and the morphological parameters of neurons are calculated firstly. Second, the improved locally linear embedding algorithm is used to reduce data dimensionality. Finally, support vector machine (SVM) algorithm is used to train and test samples. Experimental results show under certain conditions the classification of the method has good classification.

scholarly journals Handwritten Digit Recognition using Image Preprocessing and CNN

2020 ◽  
pp. 896-902
Author(s):  
Anshul Dubey ◽  
Ashley Lazarus ◽  
Dharmendra Mangal
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Accuracy Rate ◽  
Handwritten Digit Recognition ◽  
Image Preprocessing ◽  
Digit Recognition ◽  
Handwritten Digit ◽  
Multilayer Architecture ◽  
Recognition Systems ◽  
Zip Code

Handwritten digit recognition, is a technique of identifying and enlisting the recognized digit, that uses neural networks, deep learning and machine learning. The applications and demand of handwritten digit recognition systems such as zip code recognition, car number plate recognition, robotics, banks, mobile applications and numerous more, are soaring every day. It can be done through numerous approaches, but convolutional neural network is considered one of the best methods. The special neural network uses multilayer architecture for identification and classification. Although the accuracy factor can be increased, based on image preprocessing, in this paper we discuss how the accuracy of the system can be increased for better handwritten digit recognition, using convolutional neural networks, image preprocessing; binarization, resizing, rotation. The accuracy rate obtained is 99.33%.

Sign in / Sign up

Export Citation Format

Share Document