FARTrack: Fast Autoregressive Visual Tracking with High Performance

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Wang, Guijie, Lin, Tong, Bai, Yifan, Cao, Anjia, Liang, Shiyi, Zhao, Wangbo, Wei, Xing
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866908869889884160
author Wang, Guijie
Lin, Tong
Bai, Yifan
Cao, Anjia
Liang, Shiyi
Zhao, Wangbo
Wei, Xing
author_facet Wang, Guijie
Lin, Tong
Bai, Yifan
Cao, Anjia
Liang, Shiyi
Zhao, Wangbo
Wei, Xing
contents Inference speed and tracking performance are two critical evaluation metrics in the field of visual tracking. However, high-performance trackers often suffer from slow processing speeds, making them impractical for deployment on resource-constrained devices. To alleviate this issue, we propose FARTrack, a Fast Auto-Regressive Tracking framework. Since autoregression emphasizes the temporal nature of the trajectory sequence, it can maintain high performance while achieving efficient execution across various devices. FARTrack introduces Task-Specific Self-Distillation and Inter-frame Autoregressive Sparsification, designed from the perspectives of shallow-yet-accurate distillation and redundant-to-essential token optimization, respectively. Task-Specific Self-Distillation achieves model compression by distilling task-specific tokens layer by layer, enhancing the model's inference speed while avoiding suboptimal manual teacher-student layer pairs assignments. Meanwhile, Inter-frame Autoregressive Sparsification sequentially condenses multiple templates, avoiding additional runtime overhead while learning a temporally-global optimal sparsification strategy. FARTrack demonstrates outstanding speed and competitive performance. It delivers an AO of 70.6% on GOT-10k in real-time. Beyond, our fastest model achieves a speed of 343 FPS on the GPU and 121 FPS on the CPU.
format Preprint
id arxiv_https___arxiv_org_abs_2602_03214
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle FARTrack: Fast Autoregressive Visual Tracking with High Performance
Wang, Guijie
Lin, Tong
Bai, Yifan
Cao, Anjia
Liang, Shiyi
Zhao, Wangbo
Wei, Xing
Computer Vision and Pattern Recognition
Inference speed and tracking performance are two critical evaluation metrics in the field of visual tracking. However, high-performance trackers often suffer from slow processing speeds, making them impractical for deployment on resource-constrained devices. To alleviate this issue, we propose FARTrack, a Fast Auto-Regressive Tracking framework. Since autoregression emphasizes the temporal nature of the trajectory sequence, it can maintain high performance while achieving efficient execution across various devices. FARTrack introduces Task-Specific Self-Distillation and Inter-frame Autoregressive Sparsification, designed from the perspectives of shallow-yet-accurate distillation and redundant-to-essential token optimization, respectively. Task-Specific Self-Distillation achieves model compression by distilling task-specific tokens layer by layer, enhancing the model's inference speed while avoiding suboptimal manual teacher-student layer pairs assignments. Meanwhile, Inter-frame Autoregressive Sparsification sequentially condenses multiple templates, avoiding additional runtime overhead while learning a temporally-global optimal sparsification strategy. FARTrack demonstrates outstanding speed and competitive performance. It delivers an AO of 70.6% on GOT-10k in real-time. Beyond, our fastest model achieves a speed of 343 FPS on the GPU and 121 FPS on the CPU.
title FARTrack: Fast Autoregressive Visual Tracking with High Performance
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2602.03214