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

In this paper, we strive to answer the question "how to collaboratively learnconvolutional neural network (CNN)-based and vision transformer (ViT)-basedmodels by selecting and exchanging the reliable knowledge between them forsemantic segmentation?" Accordingly, we propose an online knowledgedistillation (KD) framework that can simultaneously learn compact yet effectiveCNN-based and ViT-based models with two key technical breakthroughs to takefull advantage of CNNs and ViT while compensating their limitations. Firstly,we propose heterogeneous feature distillation (HFD) to improve students'consistency in low-layer feature space by mimicking heterogeneous featuresbetween CNNs and ViT. Secondly, to facilitate the two students to learnreliable knowledge from each other, we propose bidirectional selectivedistillation (BSD) that can dynamically transfer selective knowledge. This isachieved by 1) region-wise BSD determining the directions of knowledgetransferred between the corresponding regions in the feature space and 2)pixel-wise BSD discerning which of the prediction knowledge to be transferredin the logit space. Extensive experiments on three benchmark datasetsdemonstrate that our proposed framework outperforms the state-of-the-art onlinedistillation methods by a large margin, and shows its efficacy in learningcollaboratively between ViT-based and CNN-based models.

A Good Student is Cooperative and Reliable: CNN-Transformer Collaborative Learning for Semantic Segmentation