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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Long-tailed distribution of semantic categories, which has been often ignoredin conventional methods, causes unsatisfactory performance in semanticsegmentation on tail categories. In this paper, we focus on the problem oflong-tailed semantic segmentation. Although some long-tailed recognitionmethods (e.g., re-sampling/re-weighting) have been proposed in other problems,they can probably compromise crucial contextual information and are thus hardlyadaptable to the problem of long-tailed semantic segmentation. To address thisissue, we propose MEDOE, a novel framework for long-tailed semanticsegmentation via contextual information ensemble-and-grouping. The proposedtwo-sage framework comprises a multi-expert decoder (MED) and a multi-expertoutput ensemble (MOE). Specifically, the MED includes several "experts". Basedon the pixel frequency distribution, each expert takes the dataset maskedaccording to the specific categories as input and generates contextualinformation self-adaptively for classification; The MOE adopts learnabledecision weights for the ensemble of the experts' outputs. As a model-agnosticframework, our MEDOE can be flexibly and efficiently coupled with variouspopular deep neural networks (e.g., DeepLabv3+, OCRNet, and PSPNet) to improvetheir performance in long-tailed semantic segmentation. Experimental resultsshow that the proposed framework outperforms the current methods on bothCityscapes and ADE20K datasets by up to 1.78% in mIoU and 5.89% in mAcc.

MEDOE: A Multi-Expert Decoder and Output Ensemble Framework for Long-tailed Semantic Segmentation