Abstract
Recently, more and more 3D shape datasets have become publicly available and significant results have been attained in 3D shape classification with 3D volumetric convolutional neural networks. However, the existing 3D volumetric networks have a problem with balancing model scale and classification accuracy. To address this problem, neural architecture search (NAS) was introduced into 3D shape classification tasks to search for a model satisfying both requirements. Automatically generating neural networks under NAS has attracted increasing research interest in recent years. The models learned by NAS outperform many manually designed networks in several 2D tasks like image classification, detection and semantic segmentation. In this paper, the differentiable formulation of NAS is exploited to search for several repeatable computation cells. The introduction of many light-weight designs for 3D CNNs assists in the construction of deep models with fewer parameters. The loss for the classification task along with the loss for orientation prediction are combined to guide such search. Extensive experiments are designed to evaluate candidate models on three datasets. The results demonstrate that without any pretraining, our discovered model for 3D shape classification outperforms most manually designed networks with small parameter sizes, whilst also showing that our model achieves a balance between model scale and classification accuracy.
Highlights
With the development of various machine applications like unmanned vehicles, autonomous robots and human-machine interaction, 3D computer vision tasks like 3D object classification, 3D object retrieval and 3D semantic labeling have drawn increasing attention in the past several years
RELATED WORKS we briefly introduce recent works closely related to our work, including 3D shape classification and neural architecture search in subsection
We report the 3D object classification results of our searched models and the comparison to the state-of-the-art methods
Summary
With the development of various machine applications like unmanned vehicles, autonomous robots and human-machine interaction, 3D computer vision tasks like 3D object classification, 3D object retrieval and 3D semantic labeling have drawn increasing attention in the past several years. Deep learning based methods developed rapidly and largely exceeded the performance of traditional algorithms in many recent works. For 3D shape classification, there are two mainstream CNN based methods; namely voxel based methods and multiview based methods. Wu et al [11] proposed 3D ShapeNets, which is the first attempt to extend 3D CNNs to 3D shapes. The 3D shape classification and retrieval tasks were performed on a five-layer convolutional deep belief network along with the next-best-view prediction. VoxNet [12] is another attempt to combine 3D shape classification with a shallow 3D network. Sedaghat et al [13] significantly improved the classification
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.