CNN Assignment - Milestone 1

Overview

Assignment

1. Create new Java files/classes entitled "LeNet5.java" and "MNISTCNN.java" They will be used to implement the LeNet5 CNN and the set-up and testing of LeNet5 on the MNIST data set.
2. Since you already working with the MNIST data set, importing it into your project, reading the files and displaying the images should be reused from your FFnet project. Notice that you need to modifiy the drawImage() procedure to accept a 2D array. Do all of this in your "MNISTCNN.java" file.
3. In the "LeNet5.java" file, for this milestone, implement the forward pass as documented in section II.B of the GradientBased Learning Applied to Document Recognition paper by Yann LeCun, Leon Bottou, Yoshua Bengio and Patrick Haner.
4. Set-up the weight matrices as we worked out in class through the LeNet5 architecture worksheet.
5. Ensure that in the prior part, you do not use magic numbers. Define constants that you will use to declare the arrays and that you will use later on in the code.
6. Initialize the weights as specified on page 41 of the paper.
7. Next, implement the padding (unless you send LeNet5 already padded inputs).
8. Implement C1, the first layer, ensuring you use the tanh activation function as specified on page 8 [formuli (5) and (6)] as well as Appendix A.
9. Please ensure you use structured programming principles. You should have procedures for initializing the weights, for the convolutional layers, for the pooling layers and for the classifier as necessary.
10. Implement S2, using the expression as specified on page 7, rather than MaxP from the worksheet.
11. Implement C3, using the connectivity as specified in table I.
12. Implement S4. This should be the same process as used for S2.
13. Implement C5. Be mindful of the full connectivity and how it is set up.
14. Implement F6.
15. Implement the Output layer. Please use the expression as specified on page 8 in formula (7).
16. Test your network. To test it please do the following:
    1. Comment out the initializations of the weights and instead set all of them to zero, including the bias weights. Then set the center of the filter to 1.0. For the C6, F6 and Ouput layers, set the weights to 1.0.
    2. Deactivate the activation functions for all layers, except for S2 and S4. Simply use the weighted sum of the inputs as activation.
    3. Duplicate the "displayImage" and "drawPixel" methods and change the method signatures so that they accept an array. Modify the methods so that they can be used to display the single layers of the classifier. Please limit the image width to 40.
    4. Display the following images, ensuring they are properly labeled. (Simply pass in the layer name when creating a new JFrame.) Input, C1, S2, C3, S4, C5, F6, Output. Please only display the images of the first of the feature maps and the first of the pooling layer map at each level.

Submission