ProtoFlow: Mitigating Forgetting in Class-Incremental Remote Sensing Segmentation via Low-Curvature Prototype Flow

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Wu, Jiekai, Fu, Rong, Li, Chuangqi, Zhang, Zijian, Wu, Guangxin, Zhang, Hao, Lin, Shiyin, Ni, Jianyuan, Li, Yang, Zhang, Dongxu, Gandomi, Amir H., Fong, Simon, Feng, Pengbin
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866909054471766016
author Wu, Jiekai
Fu, Rong
Li, Chuangqi
Zhang, Zijian
Wu, Guangxin
Zhang, Hao
Lin, Shiyin
Ni, Jianyuan
Li, Yang
Zhang, Dongxu
Gandomi, Amir H.
Fong, Simon
Feng, Pengbin
author_facet Wu, Jiekai
Fu, Rong
Li, Chuangqi
Zhang, Zijian
Wu, Guangxin
Zhang, Hao
Lin, Shiyin
Ni, Jianyuan
Li, Yang
Zhang, Dongxu
Gandomi, Amir H.
Fong, Simon
Feng, Pengbin
contents Remote sensing segmentation in real deployment is inherently continual: new semantic categories emerge, and acquisition conditions shift across seasons, cities, and sensors. Despite recent progress, many incremental approaches still treat training steps as isolated updates, which leaves representation drift and forgetting insufficiently controlled. We present ProtoFlow, a time-aware prototype dynamics framework that models class prototypes as trajectories and learns their evolution with an explicit temporal vector field. By jointly enforcing low-curvature motion and inter-class separation, ProtoFlow stabilizes prototype geometry throughout incremental learning. Experiments on standard class- and domain-incremental remote sensing benchmarks show consistent gains over strong baselines, including up to 1.5-2.0 points improvement in mIoUall, together with reduced forgetting. These results suggest that explicitly modeling temporal prototype evolution is a practical and interpretable strategy for robust continual remote sensing segmentation. Open-source code:https://github.com/dudududke/protoflow.
format Preprint
id arxiv_https___arxiv_org_abs_2604_03212
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle ProtoFlow: Mitigating Forgetting in Class-Incremental Remote Sensing Segmentation via Low-Curvature Prototype Flow
Wu, Jiekai
Fu, Rong
Li, Chuangqi
Zhang, Zijian
Wu, Guangxin
Zhang, Hao
Lin, Shiyin
Ni, Jianyuan
Li, Yang
Zhang, Dongxu
Gandomi, Amir H.
Fong, Simon
Feng, Pengbin
Computer Vision and Pattern Recognition
Remote sensing segmentation in real deployment is inherently continual: new semantic categories emerge, and acquisition conditions shift across seasons, cities, and sensors. Despite recent progress, many incremental approaches still treat training steps as isolated updates, which leaves representation drift and forgetting insufficiently controlled. We present ProtoFlow, a time-aware prototype dynamics framework that models class prototypes as trajectories and learns their evolution with an explicit temporal vector field. By jointly enforcing low-curvature motion and inter-class separation, ProtoFlow stabilizes prototype geometry throughout incremental learning. Experiments on standard class- and domain-incremental remote sensing benchmarks show consistent gains over strong baselines, including up to 1.5-2.0 points improvement in mIoUall, together with reduced forgetting. These results suggest that explicitly modeling temporal prototype evolution is a practical and interpretable strategy for robust continual remote sensing segmentation. Open-source code:https://github.com/dudududke/protoflow.
title ProtoFlow: Mitigating Forgetting in Class-Incremental Remote Sensing Segmentation via Low-Curvature Prototype Flow
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2604.03212