Deep-focus handheld camera captures clear 3D dental images

Before performing any major operation, medical professionals like to get a clear picture of the problem. For dental work, that could mean a 3D image of the mouth captured by a handheld camera. Such an image has the potential to provide a fast, accurate visual representation of the relevant area, but existing technologies come with limitations such as bulky hardware or motion artifacts that limit practical use.

Kwon et al. developed a deep focus light-field camera (DF-LFC) equipped with a solid immersion microlens array (siMLA) capable of producing accurate images with a large depth of field. Their camera, integrated into a 3D intraoral scanner, captured high-resolution images that could be reconstructed into a complete dental model.

“The ability of light-field cameras to capture both spatial and directional information in a single exposure, as well as their ability to be easily miniaturized, suggested an innovative method for improving 3D intraoral scanners,” said author Ki-Hun Jeong.

The major downside to LFCs is that the microlens arrays they depend on have shallow depths of field owing to their manufacturing process. In a handheld scanner, this can result in blurry images and motion artifacts. The team overcame this challenge by employing si-MLAs, which are covered with a low-index transparent elastomer.

In tests, the camera not only produced sharp 3D images but also increased contrast and reduced lens crosstalk. The authors hope to implement this technology in other related applications.

“Based on these achievements, our next steps will involve the optimization of light-field camera design and light-field image processing for advanced 3D imaging systems in clinical endoscopy or robotic surgery,” said Jeong.

Source: “Deep focus light-field camera for handheld 3D intraoral scanning using crosstalk-free solid immersion microlens arrays,” by Jae-Myeong Kwon, Sang-In Bae, Taehan Kim, Jeong Kun Kim, and Ki-Hun Jeong, APL Bioengineering (2023). The article can be accessed at https://doi.org/10.1063/5.0155862 .