ARI-AgroBench: Agri-Robustness Index (ARI) and AgroBench Benchmark

Read full text
Authors: Yassine Zarrouk
Cited by: 0

Abstract

AgroBench v1.1.0 AgroBench v1.1.0 is the software release used to generate the multi-architecture robustness experiments reported in the associated study. This release provides the complete reproducibility pipeline for evaluating agricultural object detectors under the AgroBench corruption benchmark, including: Deterministic corruption generation; Multi-architecture evaluation using YOLOv8n, YOLO11n, RT-DETR, and Faster R-CNN; Benchmark execution and evaluation orchestration; Faster R-CNN training; Corruption-specific performance evaluation and robustness aggregation; Environment and dependency specifications; Dataset configuration and reproduction instructions. This release supersedes the earlier software version used in the companion MethodsX methodology paper while preserving the previous release for version-specific reproducibility. Version: v1.1.0 Associated study: Multi-architecture robustness evaluation of agricultural object detectors using AgroBench.

Fields of Study & Subjects

Smart Agriculture and AI Advanced Neural Network Applications Adversarial Robustness in Machine Learning

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
8 March 2026
Language
EN