Machine learning for image-based multi-omics analysis of leaf veins.

The article focuses on the study of leaf vein networks in plants, emphasizing their structural and functional roles in growth, nutrient transport, and environmental adaptation. It reviews the integration of plant physiology with advanced image recognition technologies, including deep learning and machine learning (ML), to extract vein phenotypes and analyze their development under genetic and environmental influences. The article also discusses the application of multi-omics data—encompassing genomics, transcriptomics, proteomics, metabolomics, and enviromics—and ML-based multi-omics association analysis methods, both supervised and unsupervised, to elucidate the regulatory mechanisms of vein formation and improve crop productivity. Challenges in high-throughput imaging, accurate vein segmentation, and data annotation are highlighted, alongside recent technological advances such as convolutional neural networks and graph convolution networks that enhance vein network analysis. Overall, the review underscores the potential of combining multi-disciplinary approaches and artificial intelligence to advance understanding of leaf vein phenomics and its applications in agriculture and plant science.