A novel CapsNet neural network based on MobileNetV2 structure for robot image classification

Abstract

Image classification indicates that it classifies the images into a certain category according to the information in the image. Therefore, extracting image feature information is an important research content in image classification. Traditional image classification mainly uses machine learning methods to extract features. With the continuous development of deep learning, various deep learning algorithms are gradually applied to image classification. However, traditional deep learning-based image classification methods have low classification efficiency and long convergence time. The training networks are prone to over-fitting. In this paper, we present a novel CapsNet neural network based on the MobileNetV2 structure for robot image classification. Aiming at the problem that the lightweight network will sacrifice classification accuracy, the MobileNetV2 is taken as the base network architecture. CapsNet is improved by optimizing the dynamic routing algorithm to generate the feature graph. The attention module is introduced to increase the weight of the saliency feature graph learned by the convolutional layer to improve its classification accuracy. The parallel input of spatial information and channel information reduces the computation and complexity of network. Finally, the experiments are carried out in CIFAR-100 dataset. The results show that the proposed model is superior to other robot image classification models in terms of classification accuracy and robustness.

HFCC-Net: A Dual-Branch Hybrid Framework of CNN and CapsNet for Land-Use Scene Classification

“…Yousra et al [44] extract features from the VGG19 model trained on the ImageNet dataset and input them into the newly designed CapsNet to obtain recognition results. In contrast to the above approach, Zhang et al [45] input the features acquired through the CapsNet into MobileNetV2, which achieves the lightweight and accurate recognition requirements. This method is still based on CapsNet in essence, the difference is that the original input image data are replaced by the DSFM after the convolution operation; although it achieves certain image classification requirements, in the process of convolution of the image, part of the information is also lost.…”

Robust robot image classification toward cyber-physical system-based closed-loop package design evaluation

“…Robot classification for package design aims to enable robots or computers to make esthetic decisions about images of the package design in a way of imitating human vision and esthetic thinking (Zhang et al, 2022 ). The image esthetic quality assessment methods can be categorized into two main streams.…”