English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Semi-supervised learning via teacher-student network can train a modeleffectively on a few labeled samples. It enables a student model to distillknowledge from the teacher's predictions of extra unlabeled data. However, suchknowledge flow is typically unidirectional, having the performance vulnerableto the quality of teacher model. In this paper, we seek to robust 3Dreconstruction of stereo endoscopic images by proposing a novel fashion ofbidirectional learning between two learners, each of which can play both rolesof teacher and student concurrently. Specifically, we introduce twoself-supervisions, i.e., Adaptive Cross Supervision (ACS) and Adaptive ParallelSupervision (APS), to learn a dual-branch convolutional neural network. The twobranches predict two different disparity probability distributions for the sameposition, and output their expectations as disparity values. The learnedknowledge flows across branches along two directions: a cross direction(disparity guides distribution in ACS) and a parallel direction (disparityguides disparity in APS). Moreover, each branch also learns confidences todynamically refine its provided supervisions. In ACS, the predicted disparityis softened into a unimodal distribution, and the lower the confidence, thesmoother the distribution. In APS, the incorrect predictions are suppressed bylowering the weights of those with low confidence. With the adaptivebidirectional learning, the two branches enjoy well-tuned supervisions, andeventually converge on a consistent and more accurate disparity estimation. Theextensive and comprehensive experimental results on four public datasetsdemonstrate our superior performance over other state-of-the-arts with arelative decrease of averaged disparity error by at least 9.76%.

Bidirectional Semi-supervised Dual-branch CNN for Robust 3D Reconstruction of Stereo Endoscopic Images via Adaptive Cross and Parallel Supervisions