A large number of high schools were forced to close their gyms and postpone their sports seasons as a result of the COVID-19 outbreak. Several folks began doing yoga or running at their homes. Michelle Hua, a sophomore in high school, didn’t think that was good enough. An app developed by a Cranbrook Kingswood High School teen in Bloomfield Hills, Michigan, to track her travels was developed for her. It not only helps her recognize her exercises but also gives her coaching recommendations.
As a result, she’s been able to maintain her level of activity throughout the outbreak. This week, she was presented with the $75,000 George D. Yancopoulos Innovator Award in recognition of her work on the application. To receive the top award at this year’s Regeneron International Science and Engineering Fair, we have… (See the box at the bottom of this page for a complete list of the other recipients.) The Society for Science, which publishes this journal, organized the International Science and Engineering Fair (ISEF), which took place in October. This year, science competitions were held entirely online.
Michelle is a rhythmic gymnast, which means that she performs floor exercises with the help of a hoop, a ribbon, or a ball. In this exercise, gymnastics and dance are mixed to provide a unique experience. Her gym, on the other hand, was closed during COVID-19. Working on her computer at home, Michelle was disappointed with her progress and decided to quit. She wished to increase her level of physical fitness. As a result, the youngster developed an application that tracks her whereabouts at all times. In reality, it tells her whether or not she is doing things correctly.
Some movement-detection apps use bone models as the basis for movement detection, which is a common practice. They examine a video to determine which parts of the body are moving, and then they determine which regions of the body are moving, and the process continues. Michelle, on the other hand, believes that this technique is not incredibly accurate. The computer must be aware of the location of body components in each frame of the film, according to her. “It must be aware of the location of the head, shoulders, arms, legs, and feet,” she explains.
As an alternative, she chose to work with silhouettes, which are the outlines of people as a group. It is not necessary to have “knowledge about the positioning of bodily components” while employing silhouettes, according to her. A human’s outline, regardless of where the head, arms, or legs are located, must be separated from the background it is in order for it to be recognized.”
The gymnast employed a neural-net technology to design her routine, which she shared with the audience. The data on which these artificial intelligence programs are trained allows them to learn from the material they are exposed to. Michelle has used data from a number of different movement files to train her own dog. In the tens of thousands of hours of footage, people can be seen doing anything from sitting to jumping to jogging and everything in between. Each video was evaluated by her algorithm, which then produced a silhouette and learned what it was doing from the information it received.
All of your actions, from combing your hair to chewing gum, are now recognized by your computer. Additionally, it can distinguish between jumping jacks and other exercises. However, there were no movement data sets in her software that contained rhythmic gymnastics, which was a disappointment to her. As a result, the adolescent made the decision to videotape herself performing her actions. She used her own data to train her model, which she then used to test it.
Her groundbreaking approach has determined precisely what each silhouette should be doing when exercising. In the aftermath of a person’s jump, an app captures their silhouette and instructs them to lift their arms higher. According to Michelle, the feedback provided by this app assists users in modifying their posture in order to avoid any exercise-related injuries. She has developed an app with the goal of assisting individuals in exercising more effectively. It could also be used to determine how well people who are recovering from injuries are responding to their physical rehabilitation.
The adolescent worked in collaboration with Sichuan Zhong on the creation of her project. He is a computer scientist at Wayne State University in Detroit, Michigan, where he lives with his family. Michelle’s research, with the assistance of her co-authors, was published in the journal Computer Aided Geometric Design.
Michelle’s next step will be to publish her software in the Apple store, as she explains. For the time being, she and “my younger brother” are the only ones who use her app. Our health has benefited from the program’s efforts to keep us involved and active throughout the year.”