Point Cloud Acceleration by Exploiting Geometric Similarity
Cen Chen, Xiaofeng Zou, Hongen Shao, Yangfan Li, Kenli Li
Abstract
Deep learning on point clouds has attracted increasing attention for various emerging 3D computer vision applications, such as autonomous driving, robotics, and virtual reality. These applications interact with people in real-time on edge devices and thus require low latency and low energy. To accelerate the execution of deep neural networks (DNNs) on point clouds, some customized accelerators have been proposed, which achieved a significantly higher performance with reduced energy consumption than GPUs and existing DNN accelerators.
Topics & Concepts
Computer sciencePoint cloudArtificial intelligenceLatency (audio)Cloud computingDeep learningHardware accelerationAccelerationEnergy consumptionEnhanced Data Rates for GSM EvolutionRoboticsConvolutional neural networkLow latency (capital markets)Artificial neural networkDeep neural networksPoint (geometry)Edge deviceVirtual realityRobotEmbedded systemField-programmable gate arrayOperating systemComputer networkEngineeringTelecommunicationsPhysicsGeometryMathematicsElectrical engineeringClassical mechanics3D Shape Modeling and AnalysisRemote Sensing and LiDAR ApplicationsRobotics and Sensor-Based Localization