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

Training a convolutional neural network (CNN) to detect infrared smalltargets in a fully supervised manner has gained remarkable research interestsin recent years, but is highly labor expensive since a large number ofper-pixel annotations are required. To handle this problem, in this paper, wemake the first attempt to achieve infrared small target detection withpoint-level supervision. Interestingly, during the training phase supervised bypoint labels, we discover that CNNs first learn to segment a cluster of pixelsnear the targets, and then gradually converge to predict groundtruth pointlabels. Motivated by this "mapping degeneration" phenomenon, we propose a labelevolution framework named label evolution with single point supervision (LESPS)to progressively expand the point label by leveraging the intermediatepredictions of CNNs. In this way, the network predictions can finallyapproximate the updated pseudo labels, and a pixel-level target mask can beobtained to train CNNs in an end-to-end manner. We conduct extensiveexperiments with insightful visualizations to validate the effectiveness of ourmethod. Experimental results show that CNNs equipped with LESPS can wellrecover the target masks from corresponding point labels, {and can achieve over70% and 95% of their fully supervised performance in terms of pixel-levelintersection over union (IoU) and object-level probability of detection (Pd),respectively. Code is available at https://github.com/XinyiYing/LESPS.

Mapping Degeneration Meets Label Evolution: Learning Infrared Small Target Detection with Single Point Supervision