We developed a novel framework using Deep 3D Convolutional Neural Networks (3D-CNNs) termed FeatureNet to learn machining features from CAD models of mechanical parts. FeatureNet learns the distribution of complex machining feature shapes across a large 3D model data set and discovers distinguishing features that help in recognition process automatically. To train FeatureNet, a large-scale mechanical part datasets of 3D CAD models with labeled machining features is synthetically constructed.
We create our own dataset by using Solidworks API which includes 24000 models belonging to 24 classes. And we uploaded our dataset in: Dataset
.png)
We proposed a deep 3D convolutional neural network to be our recognizer. The input of recognizer is the model with only single feature. And the output is the class input feature belonging to.
We using scikit-image library to perform these two tasks. Decomposition will find different areas which are seperated from each other. Segmentation will seperate overlapping features.
- 1 : Convergence of loss function and accuracy
- 2 : Confusion matrix
- 3 : Three test cases
.png)
.png)
MAD Lab
Department of Mechanical and Aerospace Engineering
University at Buffalo, Buffalo, NY - 14260
http://madlab.eng.buffalo.edu/