Task-Oriented Semantic Feature Transmission for Multi-Task Satellite Remote Sensing over Low-SNR Channels
This research is relevant for anyone working on satellite communication or remote sensing, as it addresses a key challenge in transmitting data over noisy channels.
This paper introduces a new framework for satellite remote sensing that directly transmits semantic features instead of raw images.
It uses a lightweight channel adaptation module to compress data for bandwidth reduction and a feature restorer to recover task-relevant information after channel corruption.
Experiments show this method outperforms traditional reconstruction-based methods, especially in low-SNR conditions.
Conventional satellite remote sensing transmission follows a reconstruct-then-infer paradigm that optimizes pixel-level fidelity, creating an objective mismatch with downstream tasks such as classification and detection, especially at low SNR. This paper investigates a task-oriented framework that bypasses image reconstruction and directly transmits semantic features extracted by a multitask-pretrained backbone. A lightweight channel adaptation module (CAM) compresses feature dimensionality for bandwidth reduction, and a feature restorer recovers task-relevant structure after channel corruption. With the backbone frozen, the CAM and task-specific downstream heads are jointly optimized with task and feature-level supervision under random-SNR training. Under the adopted AWGN setting, experiments on scene classification and object detection show consistent gains over reconstruction-oriented JSCC baselines across different SNR conditions, with the largest improvements in the low-SNR regime.
This report is categorized under AITOP's "Papers" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Task-Oriented Semantic Feature Transmission.
Source attribution: curated from arXiv cs.LG (@Shuoyuan Sun, Hongyu Wang, Mugen Peng, Wenjia Xu).