DTA-Head: Dynamic Task Alignment Head for Regression and Classification in Small Object Detection

Ye, Kaiqi; Li, Qi; Yan, Yunfeng; Wang, Xianbo; Qi, Donglian

doi:10.3390/app15169060

This is an early access version, the complete PDF, HTML, and XML versions will be available soon.

Open AccessArticle

DTA-Head: Dynamic Task Alignment Head for Regression and Classification in Small Object Detection

by

Kaiqi Ye

¹,

Qi Li

²,

Yunfeng Yan

²,

Xianbo Wang

³

and

Donglian Qi

^2,*

¹

The Ocean College, Zhejiang University, Zhoushan 316021, China

²

The College of Electrical Engineering, Zhejiang University, Hangzhou 310058, China

³

The Hainan Institute of Zhejiang University, Sanya 572025, China

^*

Author to whom correspondence should be addressed.

Appl. Sci. 2025, 15(16), 9060; https://doi.org/10.3390/app15169060 (registering DOI)

Submission received: 26 June 2025 / Revised: 9 August 2025 / Accepted: 14 August 2025 / Published: 17 August 2025

Download Versions Notes

Abstract

Detecting small targets poses significant challenges due to the limited feature information and the aggregation of features in deep feature maps. Existing single-stage detectors handle classification and regression separately, leading to inconsistent predictions and potential filtering by Non-Maximum Suppression (NMS). To address these issues, we propose the Dynamic Task Alignment (DTA) Head. This novel approach comprises two branches: regression and classification. The regression branch computes offsets and masks for feature alignment adopted by Deformable ConvNets v2 (DCNv2), while the classification branch enhances feature interaction through dynamic selection. The task-decomposition module separates features for each branch. Additionally, we introduce Diverse-Scale Channel-Specific Convolution (DSCSC) to apply diverse convolutions across specific channels and exchange channel information. Our methods achieved an AP@.5 of 30.9% on the TinyPerson dataset, a 3.3% improvement over the original model’s 27.6% and outperforming other common models.

Keywords: computer vision; small object detection; dynamic task alignment; diverse-scale channel-specific convolution

Share and Cite

MDPI and ACS Style

Ye, K.; Li, Q.; Yan, Y.; Wang, X.; Qi, D. DTA-Head: Dynamic Task Alignment Head for Regression and Classification in Small Object Detection. Appl. Sci. 2025, 15, 9060. https://doi.org/10.3390/app15169060

AMA Style

Ye K, Li Q, Yan Y, Wang X, Qi D. DTA-Head: Dynamic Task Alignment Head for Regression and Classification in Small Object Detection. Applied Sciences. 2025; 15(16):9060. https://doi.org/10.3390/app15169060

Chicago/Turabian Style

Ye, Kaiqi, Qi Li, Yunfeng Yan, Xianbo Wang, and Donglian Qi. 2025. "DTA-Head: Dynamic Task Alignment Head for Regression and Classification in Small Object Detection" Applied Sciences 15, no. 16: 9060. https://doi.org/10.3390/app15169060

APA Style

Ye, K., Li, Q., Yan, Y., Wang, X., & Qi, D. (2025). DTA-Head: Dynamic Task Alignment Head for Regression and Classification in Small Object Detection. Applied Sciences, 15(16), 9060. https://doi.org/10.3390/app15169060

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Menu

DTA-Head: Dynamic Task Alignment Head for Regression and Classification in Small Object Detection

Abstract

Share and Cite

Article Metrics

Article Access Statistics

Further Information

Guidelines

MDPI Initiatives

Follow MDPI