Sütő, JózsefZihan, Syed Abrar2026-02-122026-02-122025https://hdl.handle.net/2437/404537The purpose of this research is to determine if the modified mosaic algorithm proposed by us is more effective against the YOLO object detector than the conventional mosaic augmentation technique. The efficiency of the modified mosaic algorithm will be compared against the standard mosaic method and some other commonly used augmentation techniques (such as brightness-contrast, cutout, translation, etc). The efficiency of the augmentation methods will be measured using the Mean Average Precision (mAP) metric, specifically the mAP50 and mAP50-95 values. The datasets used for this research will be three agricultural image datasets.39enObject DetectionData augmentationMachine LearningEfficiency investigation of a modified mosaic image augmentation methodInformatika::SzámítógéptudományHozzáférhető a 2022 decemberi felsőoktatási törvénymódosítás értelmében.