Generalization across agro-ecological zones and seasons

Determine whether AgroVisNet generalizes reliably across additional agro-ecological zones and cropping seasons beyond the two districts and one cropping season represented in BD-PlantDX.

Background

BD-PlantDX contains images collected from the Bogura and Nilphamari districts of Bangladesh during one winter cropping season. Although AgroVisNet transfers well to two independently collected Bangladeshi datasets, the authors note that the benchmark’s geographic and temporal coverage remains limited.

The unresolved issue is whether the model’s accuracy and learned visual representations remain reliable when pathogen prevalence, cultivars, environmental conditions, and imaging circumstances vary across other agro-ecological regions and seasons.

References

Finally, the twelve classes cover the pathologies prevalent in two districts over one cropping season, and coverage of additional agro-ecological zones and seasons remains open.

— AgroVisNet: A lightweight Convolutional Network and the BD-PlantDX Expert-Validated Benchmark for Radish, Potato and Pointed Gourd Disease Classification  (2609.10469 - Mandal et al., 9 Sep 2026) in Section 7, Discussion and Limitations