Systems and Means of Informatics

2025, Volume 35, Issue 3, pp 17-32

DEVELOPMENT OF A SMALL-OBJECT AUGMENTATION METHOD BASED ON SUPER-RESOLUTION NETWORKS

  • P. O. Arkhipov
  • S. L. Philippskih
  • M. V. Tsukanov

Abstract

The paper examines the limitations of modern data augmentation methods when applied to images captured by unmanned aerial vehicles in scenarios characterized by high object density and small object sizes. A specialized method, Contextual Small-Object Augmentation, is proposed to intelligently place visually enhanced objects into semantically relevant regions of the image while preserving spatial realism. In particular, the study focuses on a data augmentation module that utilizes super-resolution (SR) networks to improve the visual quality of small objects. For this purpose, several state-of-the-art SR neural models - RCAN, Real-ESRGAN, and SwinIR - were selected.
Their impact on the accuracy of object detection and classification was evaluated using the SSD MobileNet V2 FPNLite 320 x 320 model trained on various versions of the VisDrone benchmark dataset. The detection results were compared against a baseline model trained on the original dataset following the evaluation protocol of the COCO Evaluation Metrics. The experimental results demonstrate that incorporating high-resolution networks into the augmentation pipeline significantly improves the detection accuracy of small objects while maintaining computational efficiency.

[+] References (31)

[+] About this article