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

Annotating remote sensing images (RSIs) presents a notable challenge due toits labor-intensive nature. Semi-supervised object detection (SSOD) methodstackle this issue by generating pseudo-labels for the unlabeled data, assumingthat all classes found in the unlabeled dataset are also represented in thelabeled data. However, real-world situations introduce the possibility ofout-of-distribution (OOD) samples being mixed with in-distribution (ID) sampleswithin the unlabeled dataset. In this paper, we delve into techniques forconducting SSOD directly on uncurated unlabeled data, which is termed Open-SetSemi-Supervised Object Detection (OSSOD). Our approach commences by employinglabeled in-distribution data to dynamically construct a class-wise feature bank(CFB) that captures features specific to each class. Subsequently, we comparethe features of predicted object bounding boxes with the corresponding entriesin the CFB to calculate OOD scores. We design an adaptive threshold based onthe statistical properties of the CFB, allowing us to filter out OOD sampleseffectively. The effectiveness of our proposed method is substantiated throughextensive experiments on two widely used remote sensing object detectiondatasets: DIOR and DOTA. These experiments showcase the superior performanceand efficacy of our approach for OSSOD on RSIs.

Semi-Supervised Object Detection with Uncurated Unlabeled Data for Remote Sensing Images