End–to–end measurement technology- powered by AI
Working at the forefront of deep learning and edge computing, The Fit’s 3D body scanner measures the human body using only two photos in less than 30 seconds.
What makes The Fit stand out
Machine learning is the way we turn two photos into body data. In order to deploy machine learning models, you must first train the models. We do this with a combination of the latest and most stable frameworks. Furthermore, we use data science best practices to balance accuracy and precision in our model training and deploying processes.
By using deep neural networks in our machine learning stack, we can achieve high accuracy and add a level of privacy. Neural networks allow us to transform personal photos into unidentifiable embeddings that extract the user’s body data and remove identifiers, including facial features and clothing. In this way, we ensure user privacy and protection of personal details.
Performing calculations locally on the user’s device via edge computing means complex machine learning algorithms can be performed before communicating with our cloud services. This reduces time spent sending data back and forth. Combining cloud- and edge computing allows us to balance speed and privacy for a smooth user experience.
Overall, we strive to make our technology perform well for all user. We measure our performance and strive to improve our models metrics on 90th percentiles, instead of the averages. By looking beyond the averages, we hold ourselves to aggressive standards. We aim for performance on the long tail, because we want to serve the non-average and the average users.