FIBO_project_Module8-9 Group 2

Institute of FIeld RObotics (FIBO), King Mongkut University of Technology Thonburi (KMUTT)

Group Member

• Arthicha Srisuchinnawong zumoarthicha@gmail.com
• Chaiyaporn Boonyasathian chaiya45689@gmail.com
• Jirapat Thasingthong
• Kaewkirati Boukaew
• Kitti Pongaksorn bosskitti7983@gmail.com
• Lucksawan Yutthanakorn
• Naris Asawalertsak
• Penpat Ritprapa zefirosfenix@gmail.com
• Phurin Rangpong
• Rath Phutthitada
• Sirin Tangpoolcharoen

Description

• MNIST_data : MNIST 0-9 dataset
• data0-9compress: dataset that is used in term project (zip file; use have to unzip them)
• IP_ADDR.py : Image_Processing_And_Do_something_to_make_Dataset_be_Ready class
  • Binarization
  • Noise Removal
  • Capturing Image
  • Blur
  • Morphological Transformation
  • Perspective Transformation
  • Rotation
  • Distortion
  • Magnification
  • Font2Image
• Tenzor.py : tensorflow machine learning model class consist of 3 model
  • Neural Network (NN)
  • Convolutional Neural Network (CNN)
  • AutoEncoder (AE)
• Train.py : main program for training the agent

Suggestion

after clone this program, you should

1. change GETT_PATH and SAVE_PATH in file Train.py to your project directory
2. download and unzip dataset in data0-9compress
3. in case of you want to save CONV_PIC, you should create folder layer0 - layeri in your project directory