Kazunori Nozaki

Kazunori Nozaki

orcid

Research Focus & Metrics

Academic Career Length

14Years (2013-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
31 January 2024 in BMC Oral Health
Verified
Journal Article
Open access
PDF file
KKKazuma Kokomoto
Kazuma Kokomoto
RKRina Kariya
Rina Kariya
AMAya Muranaka
Aya Muranaka
RORena Okawa
Rena Okawa
KNKazuhiko Nakano
Kazuhiko Nakano
KNKazunori Nozaki
Kazunori Nozaki
Background: Dental age is crucial for treatment planning in pediatric and orthodontic dentistry. Dental age calculation methods can be categorized into morphological, biochemical, and radiological methods. Radiological methods are commonly used because they are non-invasive and reproducible. When radiographs are available, dental age can be calculated by evaluating the developmental stage of permanent teeth and converting it into an estimated age using a table, or by measuring the length between some landmarks such as the tooth, root, or pulp, and substituting them into regression formulas. However, these methods heavily depend on manual time-consuming processes. In this study, we proposed a novel and completely automatic dental age calculation method using panoramic radiographs and deep learning techniques. Methods: Overall, 8,023 panoramic radiographs were used as training data for Scaled-YOLOv4 to detect dental germs and mean average precision were evaluated. In total, 18,485 single-root and 16,313 multi-root dental germ images were used as training data for EfficientNetV2 M to classify the developmental stages of detected dental germs and Top-3 accuracy was evaluated since the adjacent stages of the dental germ looks similar and the many variations of the morphological structure can be observed between developmental stages. Scaled-YOLOv4 and EfficientNetV2 M were trained using cross-validation. We evaluated a single selection, a weighted average, and an expected value to convert the probability of developmental stage classification to dental age. One hundred and fifty-seven panoramic radiographs were used to compare automatic and manual human experts’ dental age calculations. Results: Dental germ detection was achieved with a mean average precision of 98.26% and dental germ classifiers for single and multi-root were achieved with a Top-3 accuracy of 98.46% and 98.36%, respectively. The mean absolute errors between the automatic and manual dental age calculations using single selection, weighted average, and expected value were 0.274, 0.261, and 0.396, respectively. The weighted average was better than the other methods and was accurate by less than one developmental stage error. Conclusion: Our study demonstrates the feasibility of automatic dental age calculation using panoramic radiographs and a two-stage deep learning approach with a clinically acceptable level of accuracy.

16 September 2021 in Scientific Reports
Verified
Journal Article
Open access
PDF file
KKKazuma Kokomoto
Kazuma Kokomoto
RORena Okawa
Rena Okawa
KNKazuhiko Nakano
Kazuhiko Nakano
KNKazunori Nozaki
Kazunori Nozaki
Dentists need experience with clinical cases to practice specialized skills. However, the need to protect patient's private information limits their ability to utilize intraoral images obtained from clinical cases. In this study, since generating realistic images could make it possible to utilize intraoral images, progressive growing of generative adversarial networks are used to generate intraoral images. A total of 35,254 intraoral images were used as training data with resolutions of 128 × 128, 256 × 256, 512 × 512, and 1024 × 1024. The results of the training datasets with and without data augmentation were compared. The Sliced Wasserstein Distance was calculated to evaluate the generated images. Next, 50 real images and 50 generated images for each resolution were randomly selected and shuffled. 12 pediatric dentists were asked to observe these images and assess whether they were real or generated. The d prime of the 1024 × 1024 images was significantly higher than that of the other resolutions. In conclusion, generated intraoral images with resolutions of 512 × 512 or lower were so realistic that the dentists could not distinguish whether they were real or generated. This implies that the generated images can be used in dental education or data augmentation for deep learning, without privacy restrictions.

1 March 2018 in Physics of Fluids
Verified
Journal Article
Closed access
PDF file
TYTsukasa Yoshinaga
Tsukasa Yoshinaga
KNKazunori Nozaki
Kazunori Nozaki
SWShigeo Wada
Shigeo Wada
The sound generation mechanisms of sibilant fricatives were investigated with experimental measurements and large-eddy simulations using a simplified vocal tract model. The vocal tract geometry was simplified to a three-dimensional rectangular channel, and differences in the geometries while pronouncing fricatives /s/ and /∫ / were expressed by shifting the position of the tongue and its constricted flow channel. Experimental results showed that the characteristic peak frequency of the fricatives decreased when the distance between the tongue and teeth increased. Numerical simulations revealed that the jet flow generated from the constriction impinged on the upper teeth wall and caused the main sound source upstream and downstream from the gap between the teeth. While magnitudes of the sound source decreased with increments of the frequency, amplitudes of the pressure downstream from the constriction increased at the peak frequencies of the corresponding tongue position. These results indicate that the sound pressures at the peak frequencies increased by acoustic resonance in the channel downstream from the constriction, and the different frequency characteristics between /s/ and /∫ / were produced by changing the constriction and the acoustic node positions inside the vocal tract.

3 December 2013 in Journal of Dental Research
Verified
Journal Article
Closed access
KNKazunori Nozaki
Kazunori Nozaki
TYT. Yoshinaga
T. Yoshinaga
SWS. Wada
S. Wada
The sibilant /s/ is produced by raising the tongue against the roof of the mouth to form a narrow constriction, which is adjusted so that the airstream emerging from it impinges on the incisors. However, the location where the sibilant sound occurs is unclear, as are the details of the mechanisms of its generation. In this study, we used a realistically shaped replica produced with a three-dimensional printer and demonstrated that turbulent flow was generated in the oral tract near the incisors and lips and that sufficiently developed turbulent flow generated a sound source up to 20,000 Hz at 333, 500, and 667 cm3/sec, which agrees with the range of physiological flow rates typical for /s/. The characteristics of the sound spectra agreed with those of the sibilant /s/ sound emitted by our control individual. Such a physical perspective could yield knowledge useful for oral surgery and speech science – for example, to predict how the generation of sibilants may be occasionally affected by orthodontic and prosthodontic treatments.

1 March 2017 in The Journal of the Acoustical Society of America
Verified
Journal Article
Open access
TYTsukasa Yoshinaga
Tsukasa Yoshinaga
KNKazunori Nozaki
Kazunori Nozaki
SWShigeo Wada
Shigeo Wada
The effects of tongue position on sound properties were investigated by using simplified models of sibilant fricatives /s/ and /ʃ/. These were constructed from medical images of a native Japanese-speaking subject who pronounced both fricative sounds. The sounds generated by the models were experimentally measured and compared with the subjects /s/ and /ʃ/. The position of tongue models altered the main peak frequency and spectral mean of the generated sound, which resulted in the reproduction of frequency characteristics of /s/ or /ʃ/ with the subjects tongue position.

19 June 2024 in IEEE Access
Verified
Journal Article
Open access
KKKazuma Kokomoto
Kazuma Kokomoto
RORena Okawa
Rena Okawa
KNKazuhiko Nakano
Kazuhiko Nakano
KNKazunori Nozaki
Kazunori Nozaki
This study pioneers the use of generative deep learning in pediatric dentistry to predict dental growth using panoramic radiography, going beyond numerical analysis and providing dynamic representations of tooth development. We employed StyleGAN-XL, a state-of-the-art generative adversarial network (GAN), to generate realistic images of dental development stages in children. Our dataset consisted of 8,092 anonymized panoramic radiographs from Osaka University Dental Hospital containing various dentition stages and conditions. By interpolating latent vectors from primary or mixed dentition images with those from permanent dentition, we generated continuous transitioning images that visually represented the progression of dental development. The performance of the StyleGAN-XL model was evaluated using Fréchet inception distance scores. Pivotal tuning inversion was used to project real images onto the model’s latent space, allowing us to effectively interpolate between current and future dental states. The resulting images showed a smooth transition from primary to permanent dentition, closely resembling the actual stages of dental development. This method represents a significant advancement in dental imaging and predictive analytics, offering a novel approach for clinicians and patients to visualize and understand dental growth. Our findings suggest broader applications for generative models in medical imaging, extending beyond traditional enhancement and modeling tasks. Our study highlights the transformative potential of GANs in medical imaging and provides a foundation for future advancements in predictive dentistry.