DAAF: Dual-stream adaptive attention fusion with distribution alignment for remote sensing object detection

Mo Zhou, Yue Zhou, Kai Song · 2026

Read DAAF: Dual-stream adaptive attention fusion with distribution alignment for remote sensing object detection by Mo Zhou, Yue Zhou, Kai Song (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/plos-pone-0353309.txt.

Description

Mo Zhou, Yue Zhou, and Kai Song built a detection system that finds objects such as bridges, vehicles, and storage tanks in satellite and aerial images. Their 2026 paper, presenting a method called DAAF, fuses two kinds of neural network: one strong on fine local detail, one strong on wide context. Naive fusion wastes that pairing because the two produce mismatched feature patterns.

DAAF first calibrates the two streams so their features align, then selects the most useful channels from each, and finally gates the blend pixel by pixel. On three public benchmarks it beat simple fusion by about two points of average precision, with large gains on hard categories like bridges, at almost no extra computing cost. Cleaner aerial detection feeds mapping, disaster checks, and land monitoring.

Catalog record

CREATORMo Zhou, Yue Zhou, Kai Song
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE78089
LICENSECC_BY
RIGHTSCC BY
SOURCEPLOS

Source: https://doi.org/10.1371/journal.pone.0353309

Browse the full library · About the Tower of Babel