Light bottle transformer based large scale point cloud classification

Abstract

With the rapid development of computer vision, point clouds technique was widely used in practical applications, such as obstacle detection, roadside detection, smart city construction, etc. However, how to efficiently identify the large scale point clouds is still an open challenge. For relieving the large computation consumption and low accuracy problem in point cloud classification, a large scale point cloud classification framework based on light bottle transformer (light-BotNet) is proposed. Firstly, the two-dimensional (2D) and three-dimensional (3D) feature values of large scale point cloud were extracted for constructing point cloud feature images, which employed the prior knowledge to normalize the point cloud features. Then, the feature images are input to the classification network, and the light-BotNet network is applied for point cloud classification. It is an interesting attempt to combine the traditional image features with the transformer network. For proving the performance of the proposed method, the large scale point cloud benchmark Oakland 3D is utilized. In the experiments, the proposed method achieved 98.1% accuracy on the Oakland 3D dataset. Compared with the other methods, it can both reduce the memory consumption and improve the classification accuracy in large scale point cloud classification.

References

[1] WeinmannM, SchmidtA, MalletC, et al.. Contextual classification of point cloud data by exploiting individual 3D neigbourhoods[C], 2015, 2: 271-278

[2] GuoY, WangH, HuQ, et al.. Deep learning for 3D point clouds: a survey[J]. IEEE transactions on pattern analysis and machine intelligence, 2020, 43(12):4338-4364

[3] WangL, MengW, XiR, et al.. 3D point cloud analysis and classification in large-scale scene based on deep learning[J]. IEEE access, 2019, 7: 55649-55658

[4] NiemeyerJ, RottensteinerF, SörgelU, et al.. Hierarchical higher order CRF for the classification of airborne lidar point clouds in urban areas[C], 2016, Hannover, Leibniz Universität Hannover: 655-662

[5] RoveriR, RahmannL, OztireliC, et al.. A network architecture for point cloud classification via automatic depth images generation[C], 2018, New York, IEEE: 4176-4184

[6] MaturanaD, SchererS. Voxnet: a 3D convolutional neural network for real-time object recognition[C], 2015, New York, IEEE: 922-928

[7] QiC R, SuH, MoK, et al.. Pointnet: deep learning on point sets for 3D classification and segmentation[C], 2017, New York, IEEE: 652-660

[8] QiC R, YiL, SuH, et al.. Pointnet++: deep hierarchical feature learning on point sets in a metric space[C], 2018, New York, Curran Associates Inc.: 30

[9] LiY, BuR, SunM, et al.. Pointcnn: convolution on x-transformed points[C], 2019, New York, Curran Associates Inc.: 1-11

[10] RamachandranP, ParmarN, VaswaniA, et al.. Stand-alone self-attention in vision models[C], 2020, New York, Curran Associates Inc.: 1-12

[11] ChenM, RadfordA, ChildR, et al.. Generative pretraining from pixels[C], 2020, New York, PMLR: 1691-1703

[12] VaswaniA, ShazeerN, ParmarN, et al.. Attention is all you need[C], 2018, New York, Curran Associates Inc.: 30

[13] SrinivasA, LinT-Y, ParmarN, et al.. Bottleneck transformers for visual recognition[C], 2021, New York, IEEE: 16519-16529

[14] BelloI, ZophB, VaswaniA, et al.. Attention augmented convolutional networks[C], 2019, New York, IEEE: 3286-3295

[15] MUNOZ D, VANDAPEL N, HEBERT M. Directional associative markov network for 3-D point cloud classification[J]. Carnegie Mellon University, 2008.

[16] LiY. Research on building classification and structure line extraction for urban vehicle borne laser point cloud data[D], 2019, Nanjing, Nanjing Normal University

[17] ElongH, HongpingW, QiC, et al.. An improved contextual classification method of point cloud[J]. Acta geodaetica et cartographica sinica, 2017, 46(3):362-370

[18] WeinmannM. Feature relevance assessment for the semantic interpretation of 3D point cloud data[C], 2013, Munich, Copernicus Publications: 2