3D Convolutional Neural Networks Initialized from Pretrained 2D Convolutional Neural Networks for Classification of Industrial Parts

Abstract

Deep learning methods have been successfully applied to image processing, mainly using 2D vision sensors. Recently, the rise of depth cameras and other similar 3D sensors has opened the field for new perception techniques. Nevertheless, 3D convolutional neural networks perform slightly worse than other 3D deep learning methods, and even worse than their 2D version. In this paper, we propose to improve 3D deep learning results by transferring the pretrained weights learned in 2D networks to their corresponding 3D version. Using an industrial object recognition context, we have analyzed different combinations of 3D convolutional networks (VGG16, ResNet, Inception ResNet, and EfficientNet), comparing the recognition accuracy. The highest accuracy is obtained with EfficientNetB0 using extrusion with an accuracy of 0.9217, which gives comparable results to state-of-the art methods. We also observed that the transfer approach enabled to improve the accuracy of the Inception ResNet 3D version up to 18% with respect to the score of the 3D approach alone.

Part Classification Prediction Based on Convolutional Neural Network

“…A convolutional neural network is a feed-forward neural network [18], which is inspired by the biological natural visual cognitive mechanism. e network of parts data can be directly input to the original data set for training, which can carry out complex preprocessing.…”

A 3d CNN Approach for Change Detection in Hr Satellite Image Time Series Based on a Pretrained 2d CNN

“…Finally, the 3D CNN extracts the 3D features that will be passed to the feature reduction module. In order to transform the convolutional layers and weights of 2D CNN to three dimensions, two techniques are considered as described in (Merino et al, 2021). Since the weights can be represented as 2D matrices, a 2D matrix can be transformed into a 3D tensor to generate the 3D weights.…”

Securing IoT Devices: A Robust and Efficient Deep Learning with a Mixed Batch Adversarial Generation Process for CAPTCHA Security Verification

“…The aforementioned observation was proven by several researchers [3][4][5][6][7][8][9][10][11][12][13][14]. DL is beneficial in other fields, including target recognition [15], speech recognition [16,17], image recognition [18][19][20], image restoration [21][22][23], audio classification [24,25], object detection [26][27][28][29][30], scene recognition [31], etc., but it has been considered "bad news" in text-based CAPTCHAs, by penetrating their security and making them vulnerable.…”