Skip to content

CNN architectures capable of extracting the annual density banding present in coral skeletons

Notifications You must be signed in to change notification settings

ainsleyrutterford/DeepC

Repository files navigation

CNN Architectures for Coral Density Analysis

Implementations of 2D SegNet (based off of Yuta Kamikawa's), 2D U-Net (based off of zhixuhao's), and 3D U-Net architectures in Keras.

Prerequisits

  • tensorflow<2
  • keras==2.2.4

These can be installed by running pip install -r requirements.txt.

If you plan on using a GPU to train, the tensorflow-gpu corresponding to the tensorflow version used is also required.

Usage

Training

$ python train.py --help

Testing

$ python test.py --help

Assessing the accuracy achieved

$ python accuracy.py --help

Predicting an entire image

$ python predict.py --help

About

CNN architectures capable of extracting the annual density banding present in coral skeletons

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published