Output files from Phenotypic trait extraction using SAM2.1 from Faba Bean Seed images.

Abstract: This dataset collection contains detailed measurements of faba bean seeds derived from high-resolution images using automated trait extraction. It includes individual seed morphometrics, shape descriptors, color features, and weight, as well as aggregate seed counts per class. These datasets are intended for research in plant phenotyping, breeding, quality assessment, and machine learning applications. They provide a comprehensive resource for analyzing seed size, shape, color, and weight relationships, enabling predictive modeling, AI/ML analyses, and statistical studies of seed traits. Dataset Files: 1. Faba_bean_Features_extraction.csv Description: Contains morphometric and shape features extracted from individual seeds using multiple segmentation and shape analysis methods, such as the Segment Anything Model (SAM), Taubin fitting, and minimum enclosing shapes. Measurements are provided in millimeters and pixels. Key Columns: · Seed number per image, Class · Morphometrics: Area-SAM(mm2), Length-SAM(mm), Width-SAM(mm), Perimeter-SAM(mm) · Shape descriptors: Eccentricity, Roundness, Circularity-SAM, Shapefactor1–4 · Spatial coordinates: centroid-0, centroid-1, bounding box (bbox-0–bbox-3) · Additional metrics: solidity, extent, Axis Major/Minor Length-SAM(pix) 2. FE_Color.csv Description: Extends the morphometric dataset with color and weight features, enabling comprehensive analysis of individual seeds. All columns from Faba_bean_Features_extraction.csv, and Additional Columns : · RGB value of Seed, color_seeds (average color and categorical label) · TGW(g) – Thousand Grain Weight of individual seeds 3. Seed Count.xlsx Description: Provides aggregate counts of seeds per class, summarizing the number of seeds observed in the dataset. Columns: · class – Seed category · seed_count – Total seeds per class