MAPLE: Multi-Path Adaptive Propagation with Level-Aware Embeddings for Hierarchical Multi-Label Image Classification

TL;DR

MAPLE is a novel framework for hierarchical multi-label image classification that combines semantic initialization, graph encoding, and adaptive fusion to improve performance on remote sensing datasets.

Contribution

It introduces a multi-path adaptive propagation method with level-aware embeddings, enhancing hierarchical modeling in multi-label classification tasks.

Findings

Achieves up to +42% improvement in few-shot regimes.

Adds only 2.6% parameter overhead.

Consistently outperforms existing methods on remote sensing datasets.

Abstract

Hierarchical multi-label classification (HMLC) is essential for modeling structured label dependencies in remote sensing. Yet existing approaches struggle in multi-path settings, where images may activate multiple taxonomic branches, leading to underuse of hierarchical information. We propose MAPLE (Multi-Path Adaptive Propagation with Level-Aware Embeddings), a framework that integrates (i) hierarchical semantic initialization from graph-aware textual descriptions, (ii) graph-based structure encoding via graph convolutional networks (GCNs), and (iii) adaptive multi-modal fusion that dynamically balances semantic priors and visual evidence. An adaptive level-aware objective automatically selects appropriate losses per hierarchy level. Evaluations on CORINE-aligned remote sensing datasets (AID, DFC-15, and MLRSNet) show consistent improvements of up to +42% in few-shot regimes while adding only 2.6% parameter overhead, demonstrating that MAPLE effectively and efficiently models hierarchical semantics for Earth observation (EO).