Answers are grounded in this paper and cite their sources.
This study aimed to develop a non-destructive method for predicting embryo presence and quantifying embryo biomass in Acer monspessulanum seeds using morphological and physical traits, combined with machine learning and statistical modeling approaches. Seeds were divided into two groups based on the presence or absence of embryos, and 26 morphological and physical traits were measured. Welch’s t-tests were used to identify traits significantly differed between seeds with and without embryos. A feedforward neural network classifier was trained on these traits to predict embryo presence. Additionally, forward stepwise regression models were constructed to identify key predictors of embryo fresh weight, dry weight, and water content in embryo‐containingseeds. Statistically significant differences (p<0.05) were observed between embryo- containing and empty seeds in traits such as seed weight, length, roundness, Hue, and floating behavior. The neural network classifier achieved 91.03% accuracy, with strong precision (0.92) and recall (0.94) for identifying embryo‐containing seeds. Regression models explained up to 56% of the variation in embryo dry weight and 50% in fresh weight, with seed weight, color parameters, and floatation traits emerging as primary predictors. Embryo water content was primarily predicted by the b* seed coat color parameter (31.6% variance), with additional contributions from buoyancy, perimeter, and post- scarification color traits, yielding a model explaining 45.5% of total variation. The integration of seed phenotyping with neural networks and multiple linear analysis provides a robust, non-destructive method for classifying embryo‐containing seed and predicting embryo development in Acer monspessulanum. This approach offers valuable applications in reforestation, seed banking, and ecological restoration initiatives.
No takes yet. Share an insight, caveat, or question.
Karimpour et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: