Zhi, ShuaifengLiu, YongxiangLi, XiangGuo, YulanIoannis Pratikakis and Florent Dupont and Maks Ovsjanikov2017-04-222017-04-222017978-3-03868-030-71997-0471https://doi.org/10.2312/3dor.20171046https://diglib.eg.org:443/handle/10.2312/3dor20171046With the rapid growth of 3D data, accurate and efficient 3D object recognition becomes a major problem. Machine learning methods have achieved the state-of-the-art performance in the area, especially for deep convolutional neural networks. However, existing network models have high computational cost and are unsuitable for real-time 3D object recognition applications. In this paper, we propose LightNet, a lightweight 3D convolutional neural network for real-time 3D object recognition. It achieves comparable accuracy to the state-of-the-art methods with a single model and extremely low computational cost. Experiments have been conducted on the ModelNet and Sydney Urban Objects datasets. It is shown that our model improves the VoxNet model by relative 17.4% and 23.1% on the ModelNet10 and ModelNet40 benchmarks with less than 67% of training parameters. It is also demonstrated that the model can be applied in real-time scenarios.I.3.5 [Computer Graphics]Computational Geometry and Object ModelingCurvesurfacesolidand object representationsI.4.8 [IMAGE PROCESSING AND COMPUTER VISION]Scene AnalysisShapeI.4.8 [IMAGE PROCESSING AND COMPUTER VISION]Scene AnalysisObject recognitionI.5.1 [PATTERN RECOGNITION]ModelsNeural nets10.2312/3dor.201710469-16