To compare the dosimetry between convex triangular fields of view (FOV) and similar dimension cylindrical FOVs of two cone-beam computed tomography (CBCT) models. Optically stimulated luminescence dosimeters (OSLDs) were placed in fiducial anatomical locations in an anthropomorphic phantom representing an adult head male for dosimetry scans. Convex triangular FOVs (100 × 80 mm/maxilla-mandible; 100 × 50 mm mandible; 100 × 50 mm/maxilla) from Veraviewepocs 3D R100 (J. Morita, Kyoto, Japan) (R100) and Veraview® X800 (J. Morita, Kyoto, Japan) (X800) and cylindrical FOVs from R100 and X800 (80 × 80 mm/maxilla-mandible; 80 × 50 mm/mandible; 80 × 50 mm/maxilla) were obtained, resulting in 12 different scan protocols. Equivalent doses for each relevant organ/tissue and the effective dose for each protocol were calculated. Mean effective doses were compared by the two-way analysis of variance (ANOVA) with Tukey's post hoc test to evaluate the effect of the FOV and device (α = 0.05). The effective doses ranged between 69 and 324 μSv for the convex triangular FOVs and 76 and 332 μSv for the cylindrical FOVs. Convex triangular FOVs from the R100 device had effective doses 2.3 to 15.3% lower than their corresponding cylindrical FOVs with similar height (p < 0.05), and that difference ranged between 8.8 and 11.8% for the X800 device (p < 0.05). Convex triangular fields of view delivered slightly lower effective doses than the cylindrical fields of view of similar dimensions in the R100 and X800 CBCT devices. Understanding the influence of the image geometry formation in effective dose allows optimization to reduce patient dose.