Search Multilayer Perceptron-Based Fusion for Efficient and Accurate Siamese Tracking

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Shen, Tianqi, Lin, Huakao, An, Ning
Format:	Preprint
Published:	2026
Subjects:	Computer Vision and Pattern Recognition Machine Learning 68T45 (Primary) 68U10, 68T10 (Secondary)
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

_version_	1866910037847310336
author	Shen, Tianqi Lin, Huakao An, Ning
author_facet	Shen, Tianqi Lin, Huakao An, Ning
contents	Siamese visual trackers have recently advanced through increasingly sophisticated fusion mechanisms built on convolutional or Transformer architectures. However, both struggle to deliver pixel-level interactions efficiently on resource-constrained hardware, leading to a persistent accuracy-efficiency imbalance. Motivated by this limitation, we redesign the Siamese neck with a simple yet effective Multilayer Perception (MLP)-based fusion module that enables pixel-level interaction with minimal structural overhead. Nevertheless, naively stacking MLP blocks introduces a new challenge: computational cost can scale quadratically with channel width. To overcome this, we construct a hierarchical search space of carefully designed MLP modules and introduce a customized relaxation strategy that enables differentiable neural architecture search (DNAS) to decouple channel-width optimization from other architectural choices. This targeted decoupling automatically balances channel width and depth, yielding a low-complexity architecture. The resulting tracker achieves state-of-the-art accuracy-efficiency trade-offs. It ranks among the top performers on four general-purpose and three aerial tracking benchmarks, while maintaining real-time performance on both resource-constrained Graphics Processing Units (GPUs) and Neural Processing Units (NPUs).
format	Preprint
id	arxiv_https___arxiv_org_abs_2603_01706
institution	arXiv
publishDate	2026
record_format	arxiv
spellingShingle	Search Multilayer Perceptron-Based Fusion for Efficient and Accurate Siamese Tracking Shen, Tianqi Lin, Huakao An, Ning Computer Vision and Pattern Recognition Machine Learning 68T45 (Primary) 68U10, 68T10 (Secondary) Siamese visual trackers have recently advanced through increasingly sophisticated fusion mechanisms built on convolutional or Transformer architectures. However, both struggle to deliver pixel-level interactions efficiently on resource-constrained hardware, leading to a persistent accuracy-efficiency imbalance. Motivated by this limitation, we redesign the Siamese neck with a simple yet effective Multilayer Perception (MLP)-based fusion module that enables pixel-level interaction with minimal structural overhead. Nevertheless, naively stacking MLP blocks introduces a new challenge: computational cost can scale quadratically with channel width. To overcome this, we construct a hierarchical search space of carefully designed MLP modules and introduce a customized relaxation strategy that enables differentiable neural architecture search (DNAS) to decouple channel-width optimization from other architectural choices. This targeted decoupling automatically balances channel width and depth, yielding a low-complexity architecture. The resulting tracker achieves state-of-the-art accuracy-efficiency trade-offs. It ranks among the top performers on four general-purpose and three aerial tracking benchmarks, while maintaining real-time performance on both resource-constrained Graphics Processing Units (GPUs) and Neural Processing Units (NPUs).
title	Search Multilayer Perceptron-Based Fusion for Efficient and Accurate Siamese Tracking
topic	Computer Vision and Pattern Recognition Machine Learning 68T45 (Primary) 68U10, 68T10 (Secondary)
url	https://arxiv.org/abs/2603.01706