Abstract: Objective: To investigate the impact of different reconstruction algorithms on image quality in ultra-high-resolution computed tomography (UHRCT) plain scans of the paranasal sinuses. Methods: Based on the clinical sinus CT protocol, five in vitro skull specimens were scanned in the high-resolution mode. Using the bone algorithm and standard algorithm, thin-layer images with 0.3125 mm thickness and spacing were reconstructed using filtered back projection (FBP), five levels of the adaptive iterative ClearView (CV) reconstruction algorithm (10%, 30%, 50%, 70%, and 90%), and five levels of the deep learning ClearInfinity (CI) reconstruction algorithm (10%, 30%, 50%, 70%, and 90%). The CT values and standard deviations (SDs) of the inferior turbinate mucosa, medial pterygoid muscle, and temporal fossa fat (background) were measured. The contrast-to-noise ratio (CNR) was calculated, and the Kruskal-Wallis H test and Mann-Whitney U test were used to analyze the objective image indicators. The image qualities obtained through the bone algorithm and the standard algorithm were evaluated subjectively, and the Kruskal-Wallis H test and Mann-Whitney U test were used for the statistical analysis of the subjective scores of the two physicians. Results: With the bone algorithm and the standard algorithms (FBP, CV, and CI), the objective indicators SD and CNR in each region of interest (ROI) showed significant differences (P < 0.05). The CI and CV algorithms decreased the SD from 90% to 10% and increased the CNR. Statistically significant differences of 30%–90% in SD and CNR were obtained using the CI, CV, and FBP algorithms (all P < 0.05). The subjective scores of the three groups of bone algorithm and standard algorithm images were significantly different between the CI 70%, CV 70%, and FBP groups (P < 0.05). The image quality of the CI 70% group was the best. Conclusions: In UHRCT plain scan imaging of the paranasal sinuses, the deep learning-based CI reconstruction algorithm can effectively reduce image noise and improve the CNR; therefore, the overall image quality is better than that of traditional FBP and CV reconstruction methods. In conclusion, CI 70% is recommended for clinical use as it provides significant advantages in visual perception and image detail presentation.