Essay Example on Recently smart systems are used in many aspects of Life









INTRODUCTION Recently smart systems are used in many aspects of life due to their effectiveness These systems use artificial intelligence such as neural networks and they can be manually trained or self trained Number recognition can be done using these neural networks by segmenting and identifying the numbers from an image even when it is written using different fonts and it can be used to improve a wide number of applications such as car number plate recognition II ARTIFICIAL NEURAL NETWORKS A Overview Artificial neural networks are inspired by the human nervous system They are made of interconnected neurons and are capable of machine learning and pattern recognition The neural networks back propagation algorithm is widely used to solve many problems B Multilayer perceptron Figure 1 Multi layer perceptron A multilayer perceptron MLP is a feed forward artificial neural network model that maps sets of input data onto a set of appropriate outputs As shown in figure 

1 A Multi layer perceptron consists of multiple layers of nodes with each layer fully connected to the next one Except for the input nodes each node is a neuron with a nonlinear activation function The network is trained using the back propagation algorithm which is a learning technique utilized by the MLP MLP is a modification of the standard linear perceptron and can distinguish data that are not linearly separable The multilayer perceptron consists of an input layer an output layer and one or more hidden layers of nonlinearly activating nodes Each node in one layer connects with a certain weight ω ij to every node in the following layer III BACK PROPAGATION ALGORITHM The Training of the systems consists of 4 steps the selection of training inputs the update of the weights the repletion and the testing The BP algorithm consists of two steps the forward propagation and the backward propagation In the forward propagation the outputs corresponding to the given inputs are calculated In the backward propagation the partial derivatives of the cost function with respect to the different parameters are propagated back through the network The network weights can then be updated using gradient based optimization algorithm The whole process is iterated until the weights are converged and the error is minimized to a global minimum

A known desired output is required by the back propagation algorithm for each input value in order to calculate the loss function gradient It is a generalization of the delta rule to multi layered made possible feed forward network by using the chain to iteratively compute gradients for each layer Back propagation requires that the activation function used by the artificial neurons or nodes be differentiable A The Network The parameters that are used for the number recognition are as below 1 Layers input 2 hidden layers output 2 64x64 inputs 3 4 output neurons based on 0 s and 1 s 4 16 neurons in the first hidden layer and 8 in the second hidden layer 5 10 desired outputs for the 10 classes corresponding to the numbers from 0 to 9 6 Initial weights are random numbers 7 Learning parameter 0 1 8 Minimum error rate 0 00001 9 Training examples for each class 8 10 Epochs 30000 11 The activation function used is the sigmoid function GUI is used to load the images train the system and test the network The code is written using Matlab The network is able to recognize the number from the image Figure 2 GUI Figure number 2 represents the GUI interface that show the parameters that are used to make the training successfully The neural network system consists of four layers the first layer is the input layer the second and the third are the Hidden layers the first hidden layer consists of 16 neurons and the second one consists of 8 neurons The last layer is the output layer and it consist of 4 neurons Before training the system the user must load the images 

A messages is displayed when the loading is done as shown in the below figure 3 and as shown in figure 4 the program displays the number of nodes per layer Figure 5 shows a message that the network displays when the training is done Figure 3 Image Loading Figure 4 Nodes per layer Figure 5 Training the network After the system is trained the user can insert the test image using the button Browse to test the network The results are shown in the figure 6 and 7 Figure 6 Tested image Figure 7 Result B Result The network was 85 able to identify from 1 to 4 objects in an image and read the numbers from 0 to 9 C Future work More improvement can be done to this network to be able to identify characters not only numbers from an image and to skip any object that is not a number or a character IV CONCLUSION Artificial networks are becoming widely used in many applications because of its efficiency and time reduction It helps in the classification of different classes thus to give almost accurate results The Back propagation algorithm plays an effective role in the classification of classes due to error correction The combination between image processing and neural networks gives great results In this project numbers are recognized using MLP and back propagation algorithm The numbers are from 0 to 9 Since the learning is supervised the outputs are compared to the desired outputs till the error drops to a minimum The project can be improved by providing more examples in order to minimize the error

Write and Proofread Your Essay
With Noplag Writing Assistance App

Plagiarism Checker

Spell Checker

Virtual Writing Assistant

Grammar Checker

Citation Assistance

Smart Online Editor

Start Writing Now

Start Writing like a PRO