Effective Data Augmentation Strategies For Small Object Detection

Author

• Tarek Aziz, Prairie View A&M University

Date of Award

8-2026

Document Type

Thesis

Degree Name

Master of Science

Department

Electrical Engineering

Abstract

Data augmentation remains the most effective method of improving object detection, particularly in scenes where small objects dominate, and the annotated datasets are small. While geometric augmentation techniques such as flipping, cropping, and photometric augmentations such as brightness, color jittering adjustments, and content-level augmentations such as mix-up and copy-paste augmentation have proved useful, current research suggests the viability of learned augmentation policies, instance-level generative augmentations, and diffusion-based augmentations. Here, we present a structured comparison of the augmentation methods on the YOLO11m detector, focusing particularly on the detection of small objects on the SODA-D dataset. We benchmarked the standard YOLO11m model with models learned with (i) conventional transforms, (ii) learned augmentation policies, (iii) diffusion-based semantic editing, and (iv) instance-level redrawing of images by employing pre-trained generative models. Through different experimental configurations, we see that dataset augmentation by instance-level augmentations always outperformed the rest, achieving the highest boosts in mean Average Precision (mAP) for object categories of small size. The research also indicated that augmentations at the object-instance level guarantee preservation of scene context and introduce visual diversity, which overcomes overfitting and improves robustness. These findings provide empirical evidence for instance-aware generative augmentations being a robust and generalizable method of detecting small objects, and leave the possibility of data-efficient training pipelines for future vision tasks.

Index Terms - Data augmentation, generative models, small object detection, yolo11m.

Committee Chair/Advisor

Suxia Cui

Committee Co-Chair

Lujun Zhai

Committee Member

Samir Abood

Publisher

Prairie View A&M University

Rights

Date of Digitization

8/14/2026

Contributing Institution

J B Coleman Library

City of Publication

Prairie View

MIME Type

Application/PDF

Recommended Citation

Aziz, T. (2026). Effective Data Augmentation Strategies For Small Object Detection. Retrieved from https://digitalcommons.pvamu.edu/pvamu-theses/1682