INTRODUCTION: Hi, This is our project implemented as a basic convolutional neural network. The feature is just identify a word on a picture being imported. The stucture of the project is separate to two parts: Convolutional Layer and Neural Network Layer. But they are intergrated in merely a CNN_NKNU.py file in the root path. You can check it out and alter the picture path if needed.
ENVIRONMENT: Recommend IDE: pycharm Environment: Anaconda venv. Needed package include "numpy", "pillow", "random", that's all.
Format of Using Classes: Convolution Layer: Same padding: = conv(PICTURE PATH, isMNIST, Layers, Times per Layer, Strides) Conventional Convolution: = FixedKernelConventionalConv(PATH, KernelSet, isMNIST, Layers, Times per Layer, Strides)
Since 2023,05,16 At this moment, there is only convolution feature separate with neural network.