Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This study benchmarks various data augmentation techniques—NAS, GAS, GMAS, and DAS—across different YOLO26 model sizes to optimize adenovirus detection in transmission electron microscopy (TEM) images. By re-annotating an adenovirus dataset to create YOLO-compatible bounding box annotations, the researchers evaluated the performance of each augmentation setup under consistent training conditions. The findings reveal significant differences in detection efficacy, highlighting the most effective data augmentation strategy for enhancing adenovirus detection accuracy.
Benchmarking reveals that the right data augmentation can dramatically enhance adenovirus detection accuracy in YOLO26 models.
This study systematically benchmarks different data augmentation setups across YOLO26 model size variants to determine the most effective setup for adenovirus detection in TEM images. The benchmarked setups include NAS, GAS, GMAS and DAS, all evaluated under identical training conditions. The adenovirus dataset, selected from the published TEM virus dataset, was re-annotated by leveraging adenovirus particle positions to generate YOLO-compatible bounding box annotations. The experimental results demonstrated the impact of the benchmarked data augmentation setups on adenovirus detection with YOLO26 and indicated the most effective data augmentation setup.