Technology from the field of computational geometry developed at the Technion enables the production of unique inhalation masks for babies

Inhalation masks adapted to the faces of babies were developed by Professor Ron Kimmel, an expert in XNUMXD data processing at the Technion Faculty of Computer Science and Professor Israel Amirav, an expert in pulmonary medicine at the "Ziv" Hospital in Safed

Parents of small children who needed inhalation, know the constant struggle and difficulty of getting small children to cooperate and accept the treatment. An innovative technology from the field of computational geometry developed at the Technion may completely change the approach to the design and production of those masks, which in the near future will become much friendlier and more efficient.

The project to develop the inhalation masks is headed by Professor Ron Kimmel, an expert in image processing and XNUMXD information and director of the Laboratory for Geometric Image Processing (GIP) at the Technion, and Professor Israel Amirav, a pediatrician specializing in pulmonary medicine, from the "Ziv" hospital in Safed.

The idea for the project was conceived by Professor Amirav, and his partner Assaf Halamish, after looking for a way to optimize the way in which toddlers inhale medicines delivered through the inhalation masks. Until now, the masks for toddlers have been designed by making the adult masks smaller, which led to a mismatch. In many cases, gaps were created between the mask in front of the toddler, through which the material emitted into the air lost its therapeutic purpose. The mismatch of the masks to the face also caused discomfort and the result was a considerable reduction in the effectiveness of the medical treatments.

"The applied problem presented by Amirav and Halamish sounded challenging and interesting and, more importantly, relevant to the field in which we are engaged, the analysis of geometric shapes," explained Professor Kimmel of his involvement in the project. Doctoral student Dan Raviv (currently researching as a postdoctoral fellow at MIT), Yaron Honan and Alon Zabirin, engineers in the Technion's laboratory for geometric processing of images, also joined the project. The idea that the researchers tested was essentially simple - the production of masks that would fit in front of toddlers in an optimal way.

"We equipped the medical staff with XNUMXD scanners that we developed in the lab and sent them to collect facial geometry (that is, a XNUMXD photograph of the facial surface) of babies in clinics," Professor Kimmel explained. Amirav and Halamish collected hundreds of examples that characterize the geometry of children from infancy to age four. At this point Raviv entered the picture and classified the hundreds of photographs into three characteristic groups, based on computational methods from the field of metric geometry.

"The three-dimensional structure of each baby's face is unique and different from the others. This, by the way, is the reason that facial recognition systems improve their performance if they use XNUMXD information alongside regular photos." Professor Kimmel explained. "However, in the range of possible changes of the three-dimensional structure of the babies' faces, we found that it is possible to characterize three main groups - this process is called clustering - the division of the information into clusters of photographs that form groups where the changes are small to the naked eye. For each of the groups we chose the 'median child' - the child's face that optimally represents the group to which he belongs, as a model according to which we will design the mask that should fit each group. We computationally verified the matching of the representative mask to his face despite the changes. After careful matching and analysis of the results, we 'printed' the three median faces using a XNUMXD printer, and customized masks were designed for them to make sure that even outliers in each group receive a good response. We printed in XNUMXD the face of the most unusual child in each group, we also printed the trachea, which was reconstructed from a computed tomography (CT) image, and thus assembled a perfect model of the respiratory system. In this way, we could test the quality of the seal obtained with the new mask on the face of a toddler, even if he was the exception in the group, and compare it to the seal obtained with the old masks."

Professor Amirav: "We decided to measure the faces of as many babies as possible to get a real idea of ​​the shape and size of the faces. For this purpose, we went to kindergartens and asked volunteer parents to agree to let us photograph the toddlers. Within a few seconds a 300D image of the face was taken and saved on the computer. About XNUMX babies were photographed this way. The camera and the technology developed at the faculty allowed us to do this in a simple and friendly way."

The highlight of the project was the possibility to insert a pacifier into the baby's mouth while putting on the mask. The pacifier enters the mask through a special slot and the baby, using the pacifier, holds the mask close to his face. "The act of sucking calms the baby and allows the treatment to be given through another hole in the mask through which the medicinal spray (aerosol) is injected which the baby breathes through the nose," according to Professor Kimmel: "The pacifier calms the baby and also allows the medicine to be injected when the mask is completely sealed."

"The masks have already been registered by the FDA (American Drug Administration) and are expected to be produced for commercial use. I hope that very soon we will see the masks in use around the world", concludes Professor Amirav. And Professor Kimmel adds that "the effectiveness of the new masks has been proven to be superior in every parameter to the existing masks on the market. And as with any medical development - the satisfaction from this is tremendous."