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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Efficient road and building footprint extraction from satellite images arepredominant in many remote sensing applications. However, precise segmentationmap extraction is quite challenging due to the diverse building structurescamouflaged by trees, similar spectral responses between the roads andbuildings, and occlusions by heterogeneous traffic over the roads. Existingconvolutional neural network (CNN)-based methods focus on either enrichedspatial semantics learning for the building extraction or the fine-grained roadtopology extraction. The profound semantic information loss due to thetraditional pooling mechanisms in CNN generates fragmented and disconnectedroad maps and poorly segmented boundaries for the densely spaced smallbuildings in complex surroundings. In this paper, we propose a novelattention-aware segmentation framework, Multi-Scale Supervised DilatedMultiple-Path Attention Network (MSSDMPA-Net), equipped with two new modulesDynamic Attention Map Guided Index Pooling (DAMIP) and Dynamic Attention MapGuided Spatial and Channel Attention (DAMSCA) to precisely extract the buildingfootprints and road maps from remotely sensed images. DAMIP mines the salientfeatures by employing a novel index pooling mechanism to retain importantgeometric information. On the other hand, DAMSCA simultaneously extracts themulti-scale spatial and spectral features. Besides, using dilated convolutionand multi-scale deep supervision in optimizing MSSDMPA-Net helps achievestellar performance. Experimental results over multiple benchmark building androad extraction datasets, ensures MSSDMPA-Net as the state-of-the-art (SOTA)method for building and road extraction.

MultiScale Probability Map guided Index Pooling with Attention-based learning for Road and Building Segmentation