[论文解读] CapsNet comparative performance evaluation for image classification

Rinat Mukhometzianov, Juan Antonio Cabrera Carrillo|arXiv (Cornell University)|May 28, 2018

Advanced Neural Network Applications参考文献 22被引用 79

一句话总结

本文在四个数据集上将 CapsNet 与 Fisherfaces、LeNet、ResNet 进行对比评估，结果显示 CapsNet 需要显著的计算资源且平均性能不及其他方法，尽管在获得更强大资源和改进的体系结构后可能具有潜力。

ABSTRACT

Image classification has become one of the main tasks in the field of computer vision technologies. In this context, a recent algorithm called CapsNet that implements an approach based on activity vectors and dynamic routing between capsules may overcome some of the limitations of the current state of the art artificial neural networks (ANN) classifiers, such as convolutional neural networks (CNN). In this paper, we evaluated the performance of the CapsNet algorithm in comparison with three well-known classifiers (Fisher-faces, LeNet, and ResNet). We tested the classification accuracy on four datasets with a different number of instances and classes, including images of faces, traffic signs, and everyday objects. The evaluation results show that even for simple architectures, training the CapsNet algorithm requires significant computational resources and its classification performance falls below the average accuracy values of the other three classifiers. However, we argue that CapsNet seems to be a promising new technique for image classification, and further experiments using more robust computation resources and re-fined CapsNet architectures may produce better outcomes.

研究动机与目标

• 在多样化图像数据集上评估 CapsNet 相对于既有分类器的性能。
• 研究训练 CapsNet 相对于传统基于 CNN 的模型的计算需求。
• 提供关于 CapsNet 潜在优势及未来研究方向的见解。

提出的方法

• 在四个图像数据集（人脸、交通标志、物体）上将 CapsNet 与 Fisherfaces、LeNet、ResNet 进行比较。
• 衡量分类准确率和训练资源需求。
• 分析 CapsNet 在简单架构和资源约束下的伸缩性。

实验结果

研究问题

• RQ1 在多个数据集上，CapsNet 的分类准确率与 Fisherfaces、LeNet、ResNet 的比较如何？
• RQ2 相对于其他分类器，训练 CapsNet 的计算资源需求是多少？
• RQ3 简单设计的 CapsNet 架构是否能达到具有竞争力的性能，需要哪些改进以获得更好结果？

主要发现

• CapsNet 的训练需要显著的计算资源。
• 在所测试的数据集上，CapsNet 的平均分类准确率低于另外三种分类器。
• CapsNet 在图像分类方面显示出潜力，若拥有更强大的资源和精炼的架构，可能带来改进。