Skip to content

Latest commit

 

History

History
26 lines (22 loc) · 1.47 KB

myreadme.md

File metadata and controls

26 lines (22 loc) · 1.47 KB

Final Project: Classification and Detection with Convolutional Neural Networks

The goal of this project is to design a digit detection and recognition system which takes in a single image and returns any sequence of digits visible in the image by using convolution neural networks.

Getting Started

Please follow the instructions on how to get the code running. First, download the all project files to your local machine. Second, get the corrected environment installed, tensorflow-gpu=1.12.0, python=3.6.7, keras=2.2.4 are used in the project.

  1. Download the all project files from gradescope into your testing directory
  2. Download the vgg model from GT Box or Dropbox
  3. Put the vgg model under your testing directoy(for example, ./cv_final_proj)
  4. cd ./your_testing_directoy, where all projects files are stored
  5. conda env create -f mycv_proj.yml
  6. conda activate cv_proj
  7. python run.py

Project folder structure

  • Test_images folder - contains all test images for this project, 1-6.png
  • graded_images folder - contains processed images
  • digit_detector.ipynb - my own designed CNN model implemented in tensorflow
  • vgg16_classifier.ipynb - VGG16 pre-trained model implemented in Keras
  • data_process.py
  • cnn_classifier.py - CNN model inference file
  • vgg_classfier.py - VGG16 inference file
  • run.py - main file to run the classifier