Hierarchical Graph Feature Enhancement with Adaptive Frequency Modulation for Visual Recognition

Convolutional neural networks (CNNs) havedemonstrated strong performance in visual recognition tasks,but their inherent reliance on regular grid structures limitstheir capacity to model complex topological relationships andnon-local semantics within images. To address this limita tion, we propose the hierarchical graph feature enhancement(HGFE), a novel framework that integrates graph-based rea soning into CNNs to enhance both structural awareness andfeature representation. HGFE builds two complementary levelsof graph structures: intra-window graph convolution to cap ture local spatial dependencies and inter-window supernodeinteractions to model global semantic relationships. Moreover,we introduce an adaptive frequency modulation module thatdynamically balances low-frequency and high-frequency signalpropagation, preserving critical edge and texture informationwhile mitigating over-smoothing. The proposed HGFE moduleis lightweight, end-to-end trainable, and can be seamlesslyintegrated into standard CNN backbone networks. Extensiveexperiments on CIFAR-100 (classification), PASCAL VOC,and VisDrone (detection), as well as CrackSeg and CarParts(segmentation), validated the effectiveness of the HGFE inimproving structural representation and enhancing overallrecognition performance.