The application of deep learning technology in smart agriculture: Lightweight apple leaf disease detection model | International Journal for Simulation and Multidisciplinary Design Optimization (IJSMDO)

Open Access

Table 3

Ablation experiment results.

Model	Improved items				mAP@0.5/%	FPS	IoU mean	single frame time/ms	missed detection rate/%

	EdgeNeXt	C2f-SC	DySample	MPDIOU
A	×	×	×	×	82.34	82	75.25	8.24	13.59
B	√	×	×	×	86.25	117	76.82	8.06	11.76
C	√	√	×	×	91.67	121	79.19	8.27	11.04
D	√	√	√	×	92.91	124	81.33	8.28	9.85
E	√	√	√	√	95.54	124	84.52	8.35	8.82

Note: “×” means the experimental group does not include the improvement item, while “√” means it does. Models without MPDIoU loss use CIoU loss by default. FPS tests are based on the NVIDIA RTX 3080Ti GPU.

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.