View Code? Open in Web Editor
NEW
This project forked from chukhinai/comp_visision
SFEDU computer vision course
License: MIT License
Python 1.07%
Jupyter Notebook 98.84%
C++ 0.09%
comp_visision's Introduction
- Put the train data and test data folder in the Data folder
- First run the data.py file to load the data and preprocess it.
- Then run the train.py file to train a Convolutional Neural Network, make prediction and store the predicted images.
- The predicted mask images can be found in the Data/output folder.
- The images are converted into arrays and stored.
- A Convolutional Neural Network(CNN) having 5 convolution2d block of filter values 64, 32 and 1. Each layer is seperated by an activation layer(Relu).
- The CNN is trained with the training data and then made to predict the output(mask images) for the testing data.
- The output is saved back as an image in the Data/output folder.
comp_visision's People
Contributors