平面设计网站有哪些比较好如何建立本站站点

MVS 无监督学习

问题描述

视角遮挡、光照变化以及光度噪声

损失函数

光度损失 (Learning unsupervised multi-view stereopsis via robust photometric consistency)

跨视图损失 (MVS2: Deep unsupervised multi view stereo with multi-view symmetry)

多阶段深度一致 (DS-MVSNet: Unsupervised multi-view stereo via depth synthesis)

额外约束

法向量一致 (M3VSNet: Unsupervised multi-metric multi-view stereo network)

神经体渲染一致性 (RC-MVSNet: Unsupervised multi-view stereo with neural rendering)

语义一致性 (Self-supervised multi view stereo via effective co-segmentation and data-augmentation)

训练策略

知识蒸馏 KD-MVS

双对比学习 CL-MVSNet: Unsupervised multi-view stereo with dual-level contrastive learning