RiMS-FiLM: Non-Destructive Multimodal Screening for Fixed-Budget Re-Inspection Prioritization of Maize Kernels.
- 1Business School, Beijing Wuzi University, Beijing 101149, China.
- 2Beijing Key Laboratory of Smart Supply Chain Application and Risk Management, Beijing Wuzi University, Beijing 101149, China.
- 3National Engineering Research Center for Agri-Product Quality Traceability, Beijing Technology and Business University, Beijing 100048, China.
Abstract
Rapid, non-destructive prioritization of maize kernels can help allocate limited re-inspection capacity in post-harvest quality control. Existing deep learning approaches mainly optimize overall classification accuracy and do not explicitly address which samples should be inspected first when verification resources are constrained. This study proposes RiMS-FiLM, an end-to-end multimodal screening framework that integrates kernel RGB appearance with lightweight physical attributes (weight and size) to generate re-inspection priority scores. Feature-wise Linear Modulation adaptively conditions visual representations on physical measurements, while a prototype-guided module organizes fused embeddings into class-structured operational quality-priority regions. The risk tiers are derived from the original GrainSet-Maize defect categories and are used as screening-priority labels rather than laboratory-validated toxicological endpoints. Internal evaluation on the stratified split of the single public GrainSet-Maize dataset over five random seeds yielded a Macro-F1 of 0.9917 ± 0.0027 and captured 94.65 ± 0.21% of high-priority samples within a 20% re-inspection budget. It also reduced severe high-to-low priority errors compared with representative fusion and selective-screening baselines. These results demonstrate the potential of low-cost multimodal sensing for resource-constrained, non-destructive grain quality screening and re-inspection planning.
Related
- DS5: An Open-Source Framework for Standardized High-Throughput Drug Screening Data Storage, Analysis, and Drug Prioritization
- CytosMPS: an automated multimodal pooled screening platform for protein, morphology and transcriptomics
- STEER-seq: a recombinase based platform for pooled multimodal single-cell perturbation screening
- Integrating socioeconomic context with multimodal EEG data for improved ADHD risk screening
- The enlightened entomologist: fast, non-destructive whole-arthropod clearing for three-dimensional imaging
BibTeX
@article{bi2026rimsfilm,
title = {{RiMS-FiLM: Non-Destructive Multimodal Screening for Fixed-Budget Re-Inspection Prioritization of Maize Kernels.}},
author = {Mingwen Bi and Ziqian Zhang and Xiaobo Yang and Feng Li and Ge Gao and Jianlei Kong and Min Zuo and Qingchuan Zhang},
journal = {Foods (Basel, Switzerland)},
year = {2026},
month = sep,
doi = {10.3390/foods15193400},
url = {https://europepmc.org/article/MED/42883507},
note = {CC BY},
}
RIS
TY - JOUR
TI - RiMS-FiLM: Non-Destructive Multimodal Screening for Fixed-Budget Re-Inspection Prioritization of Maize Kernels.
AU - Mingwen Bi
AU - Ziqian Zhang
AU - Xiaobo Yang
AU - Feng Li
AU - Ge Gao
AU - Jianlei Kong
AU - Min Zuo
AU - Qingchuan Zhang
PY - 2026
DA - 2026/09/24
JO - Foods (Basel, Switzerland)
DO - 10.3390/foods15193400
UR - https://europepmc.org/article/MED/42883507
AB - Rapid, non-destructive prioritization of maize kernels can help allocate limited re-inspection capacity in post-harvest quality control. Existing deep learning approaches mainly optimize overall classification accuracy and do not explicitly address which samples should be inspected first when verification resources are constrained. This study proposes RiMS-FiLM, an end-to-end multimodal screening framework that integrates kernel RGB appearance with lightweight physical attributes (weight and size) to generate re-inspection priority scores. Feature-wise Linear Modulation adaptively conditions visual representations on physical measurements, while a prototype-guided module organizes fused embeddings into class-structured operational quality-priority regions. The risk tiers are derived from the original GrainSet-Maize defect categories and are used as screening-priority labels rather than laboratory-validated toxicological endpoints. Internal evaluation on the stratified split of the single public GrainSet-Maize dataset over five random seeds yielded a Macro-F1 of 0.9917 ± 0.0027 and captured 94.65 ± 0.21% of high-priority samples within a 20% re-inspection budget. It also reduced severe high-to-low priority errors compared with representative fusion and selective-screening baselines. These results demonstrate the potential of low-cost multimodal sensing for resource-constrained, non-destructive grain quality screening and re-inspection planning.
SN - 2304-8158
ER -