Semi-supervised learning via teacher-student network can train a model effectively on a few labeled samples. It enables a student model to distill knowledge from the teacher's predictions of extra unlabeled data. However, such knowledge flow is typically unidirectional, having the performance vulnerable to the quality of teacher model. In this paper, we seek to robust 3D reconstruction of stereo endoscopic images by proposing a novel fashion of bidirectional learning between two learners, each of which can play both roles of teacher and student concurrently. Specifically, we introduce two self-supervisions, i.e., Adaptive Cross Supervision (ACS) and Adaptive Parallel Supervision (APS), to learn a dual-branch convolutional neural network. The two branches predict two different disparity probability distributions for the same position, and output their expectations as disparity values. The learned knowledge flows across branches along two directions: a cross direction (disparity guides distribution in ACS) and a parallel direction (disparity guides disparity in APS). Moreover, each branch also learns confidences to dynamically refine its provided supervisions. In ACS, the predicted disparity is softened into a unimodal distribution, and the lower the confidence, the smoother the distribution. In APS, the incorrect predictions are suppressed by lowering the weights of those with low confidence. With the adaptive bidirectional learning, the two branches enjoy well-tuned supervisions, and eventually converge on a consistent and more accurate disparity estimation. The extensive and comprehensive experimental results on four public datasets demonstrate our superior performance over other state-of-the-arts with a relative decrease of averaged disparity error by at least 9.76%.

翻译：基于教师-学生网络的半监督学习能够在少量标注样本上有效训练模型，使学生模型从教师对额外未标注数据的预测中蒸馏知识。然而，这种知识流通常是单向的，导致模型性能易受教师模型质量的影响。本文通过提出一种新型的双向学习方法，在两个学习者之间实现双向学习（每个学习者可同时扮演教师与学生角色），旨在鲁棒地实现立体内窥镜图像的三维重建。具体而言，我们引入两种自监督机制，即自适应交叉监督（ACS）和自适应并行监督（APS），以学习一个双分支卷积神经网络。两个分支对同一位置预测两种不同的视差概率分布，并输出其期望值作为视差值。学习到的知识沿两个方向在分支间流动：交叉方向（视差指导ACS中的分布）和并行方向（视差指导APS中的视差）。此外，每个分支还学习置信度以动态优化其提供的监督。在ACS中，预测的视差被软化为一组单峰分布，置信度越低则分布越平滑；在APS中，通过降低低置信度样本的权重来抑制错误预测。借助自适应双向学习，两个分支获得精心调优的监督，最终收敛于一致且更准确的视差估计。在四个公开数据集上的广泛实验表明，我们的方法显著优于其他现有技术，平均视差误差相对降低至少9.76%。