The sowing properties of seed material are fundamental to the reproductive success of crops and directly influence agricultural productivity and food security. Corn, as a globally significant crop, requires innovative technologies to improve seed quality and performance. This study investigated a two-stage laboratory treatment sequence for maize seeds, combining morphological fractionation with LED irradiation. The objective was to characterise and compare laboratory responses of Large Flat and Large Rounded maize-seed fractions under selected red, blue, and combined red-blue irradiation conditions. The proposed methodology included size-based separation of seeds using sieves with complex geometric holes, followed by treatment with electromagnetic radiation in the red (660 nm) and blue (460 nm) spectral ranges, individually and in combination. Large Flat and Large Rounded fractions of the DN Garant maize hybrid were obtained by sieve calibration and exposed to LED irradiation for 10 or 30 min. Seed geometry, spectral reflectance, germination energy (first-count germination), seed germination (final normal germination), and early seedling growth were evaluated under controlled laboratory conditions. The available laboratory summary values showed differences among the investigated spectrum × exposure-time × seed-fraction combinations. The highest observed first-count germination (97%) and final normal germination (95%) occurred for the Large Rounded fraction exposed to combined 660+460 nm irradiation for 30 min, compared with 83% and 85%, respectively, in the untreated reference. These values identify the highest-performing condition within the tested experimental range but do not establish a statistically confirmed optimum. Because the original replicate-level records required for verified inferential reanalysis were not available for the present revision, reconstructed replicate observations were not used as experimental evidence, and no claims of statistical significance are made from reconstructed data. A reconstructed replicate configuration was therefore used only as a sensitivity analysis to assess the robustness of the qualitative interpretation; its inferential outputs are not presented as independently verified experimental statistics. The experiment also did not include an irradiated unsorted-seed comparator, field yield measurements, biochemical assays, or industrial-scale validation. Accordingly, the conclusions are limited to descriptive laboratory differences among the investigated morphological fractions and treatment variants.