3DGradCAM

This is a repository for 3D Gradient weighted Class Activation Maps for Manufacturability Analysis. For more details, please see the paper: Learning Localized Geometric Features Using 3D-CNN: An Application to Manufacturability Analysis of Drilled Holes
Aditya Balu, Sambit Ghadai, Kin Gwn Lore, Gavin Young, Adarsh Krishnamurthy, Soumik Sarkar

Demo: Adam Lab

Packages needed

Keras 2.0.5, Tensorflow 1.2

Usage

You can either import the GradCAM function:

from GradCAM import prepareGradCAM, GradCAM
activationFunction=prepareGradCAM(cnnModel, layerIdx, nbClasses)

activationFunction is a Keras Function for computing the class activation map

License

BSD

Future work

The following codes shall be made available soon:

1. Data Generation using ACIS solid modeling kernel
2. GPU Accelerated Voxelization: Available now @ https://github.com/idealab-isu/GPView
3. Training
4. GPU accelerated Volume Rendering of GradCAM output