Automated Synthetic Data for Computer Vision: Blender-COCO Pipeline Enhancing Yolov8

• Victor Roza Souza AKCIT / UFAM https://orcid.org/0009-0004-2738-358X
• Djonathan Caua Fritzen Christ UTFPR https://orcid.org/0009-0006-3801-6288
• Igor Scaliante Wiese UTFPR https://orcid.org/0000-0001-9943-5570
• Thiago França Naves UTFPR https://orcid.org/0000-0002-3152-1197
• Telma Woerle de Lima Soares UFG https://orcid.org/0000-0002-4927-2221
• Anderson da Silva Soares UFG https://orcid.org/0000-0002-2967-6077
• Luciana de Oliverira Berretta UFG https://orcid.org/0000-0003-1857-8989

Resumo

This paper introduces an automated and reproducible pipeline for the generation of synthetic datasets with COCO-format annotations, designed to support object detection and segmentation tasks. The proposed framework combines controlled video acquisition of real-world objects, volumetric reconstruction via the SVRaster algorithm, automated mesh postprocessing in Blender, and synthetic rendering with systematically varied lighting and backgrounds. Annotation generation is fully automated through the custom YOLO FOTO plugin, which orchestrates virtual camera placement and metadata export. The resulting synthetic images are employed to train the YOLOv8n model. The study is conducted under the Design Science Research (DSR) methodology, ensuring methodological rigor and reproducibility. Experimental results demonstrate substantial improvements in model performance when synthetic data are incorporated into training, underscoring their effectiveness in enhancing dataset diversity and robustness. These findings highlight the strategic value of synthetic data in scenarios where manual annotation is costly, time-consuming, or otherwise constrained.

Palavras-chave: Training, YOLO, Three-dimensional displays, Annotations, Pipelines, Manuals, Streaming media, Rendering (computer graphics), Synthetic data, Videos, 3D reconstruction