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

Deep learning models have witnessed depth and pose estimation framework onunannotated datasets as a effective pathway to succeed in endoscopicnavigation. Most current techniques are dedicated to developing more advancedneural networks to improve the accuracy. However, existing methods ignore thespecial properties of endoscopic images, resulting in an inability to fullyunleash the power of neural networks. In this study, we conduct a detailanalysis of the properties of endoscopic images and improve the compatibilityof images and neural networks, to unleash the power of current neural networks.First, we introcude the Mask Image Modelling (MIM) module, which inputs partialimage information instead of complete image information, allowing the networkto recover global information from partial pixel information. This enhances thenetwork' s ability to perceive global information and alleviates the phenomenonof local overfitting in convolutional neural networks due to local artifacts.Second, we propose a lightweight neural network to enhance the endoscopicimages, to explicitly improve the compatibility between images and neuralnetworks. Extensive experiments are conducted on the three public datasets andone inhouse dataset, and the proposed modules improve baselines by a largemargin. Furthermore, the enhanced images we proposed, which have higher networkcompatibility, can serve as an effective data augmentation method and they areable to extract more stable feature points in traditional feature pointmatching tasks and achieve outstanding performance.

Unleashing the Power of Depth and Pose Estimation Neural Networks by Designing Compatible Endoscopic Images