Chen Y.-CChao C.-HLiu C.-LShih K.-THOMER H. CHEN2022-04-252022-04-252021https://www.scopus.com/inward/record.uri?eid=2-s2.0-85124238418&doi=10.1109%2fAIVR52153.2021.00016&partnerID=40&md5=c962587d6d282fbbdefc62361c08094chttps://scholars.lib.ntu.edu.tw/handle/123456789/607101In this paper, we focus on the speedup of a learning-based light field synthesis pipeline. The pipeline involves a disparity estimation neural network and a light field blending component. The former achieves high speed performance through the use of feature extraction and multi-stage disparity refinement, while the latter warps and merges coarse light fields generated from the left and right disparity maps in a novel and efficient way. The pipeline can produce a full light field in less than 1/10 of a second, while retaining fairly reasonable image quality. The model itself has a very low parameter count, which is ideal for devices with limited computational power. ? 2021 IEEECNNDeep learningLight field reconstructionLight field synthesisStereo visionImage reconstructionPipelinesStereo image processingDisparity estimationsField synthesisLight fieldsNeural-networksSpeed upStereoimages[SDGs]SDG7conference paper10.1109/AIVR52153.2021.000162-s2.0-85124238418