One-Time Training for All Grains: Open-Set Grain Recognition and Quantitative Analysis
2026-08-10 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors developed GROW, a system to identify and analyze different types of grains without needing to retrain the model when new grain varieties are added. It first finds grains in images and then uses a combination of visual and shape features to recognize them, storing this info in a database called GrainBank. GROW can quickly update its variety list by adding new descriptors rather than retraining the whole model. Their tests showed GROW works well across different grain amounts and backgrounds, and it is much faster to update than traditional methods. Overall, the authors provide a flexible and efficient tool for grain recognition and measurement.
crop breedinggrain variety recognitionopen set recognitionclass-agnostic localizationvisual embeddingsmorphological descriptorsinstance countingphenotypic analysisretrieval-based recognitionmodel retraining
Authors
Qihe Su, Mengyu Sun, Yuxi Ke, Zhuoyan Jiang, Wanneng Yang, Chenglong Huang, Ziyuan Yang
Abstract
Advances in crop breeding have introduced an increasing number of grain varieties, creating a growing demand for efficient variety recognition and quantitative analysis. However, existing methods are typically trained on a fixed variety set, and incorporating newly introduced varieties requires additional data collection and model retraining. To address this limitation, we propose GROW, a framework for Grain Recognition and quantitative analysis in Open sets Without retraining. GROW first performs class-agnostic grain localization, converting mixed-grain images into individual instances for variety-wise counting and phenotypic measurement. It then combines visual embeddings and morphological descriptors into fused grain descriptors stored in an extensible GrainBank. Query grains are recognized through rank-similarity weighted top-k retrieval, and newly introduced varieties are incorporated by appending their descriptors without updating the deployed models. Extensive experiments under progressive variety expansion, varying grain densities, and background domain shifts demonstrate the scalability, robustness, and adaptability of GROW. Compared with joint retraining, GROW reduced the average category-registration time from 4153 s to only 39 s while maintaining competitive recognition performance. These results demonstrate that GROW provides an efficient and maintainable solution for extensible grain recognition, counting, and phenotypic analysis without repeated model retraining.