Hinge Health has filed a patent for an apparatus that can receive raw data from an external source and generate a three-dimensional mesh based on the data. The apparatus includes a communications interface, memory storage unit, pre-processing engine, neural network engine, and mesh creator engine. The invention aims to outline objects and represent points on them using a coarse segmentation map and joint heatmap. GlobalData’s report on Hinge Health gives a 360-degree view of the company including its patenting strategy. Buy the report here.
According to GlobalData’s company profile on Hinge Health, AI-assisted medical imaging was a key innovation area identified from patents. Hinge Health's grant share as of September 2023 was 12%. Grant share is based on the ratio of number of grants to total number of patents.
The patent is filed for an apparatus to generate 3d mesh
A recently filed patent (Publication Number: US20230306686A1) describes an apparatus and method for generating a three-dimensional mesh of an object, specifically a person, using raw data received from an external source. The apparatus includes a communications interface to receive the raw data, a memory storage unit to store the raw data, a pre-processing engine to generate a coarse segmentation map and a joint heatmap from the raw data, a neural network engine to generate two-dimensional maps, and a mesh creator engine to generate a three-dimensional mesh based on the two-dimensional maps.
The raw data received by the apparatus can be a two-dimensional image, such as an RGB image. The pre-processing engine generates a coarse segmentation map to outline the person and a joint heatmap to represent a joint on the person. The neural network engine uses these maps, along with the raw data, to generate a plurality of two-dimensional maps. These maps include a fine segmentation map, a distance map, a thickness map, and various color maps. Additionally, the neural network engine removes lighting and shadow data from the raw data.
The mesh creator engine then utilizes the two-dimensional maps to generate a three-dimensional mesh of the person. The mesh is created within the fine segmentation map and pixels outside of the segmentation are discarded. The method for generating the three-dimensional mesh follows a similar process, involving receiving raw data, generating segmentation maps and joint heatmaps, applying a neural network, and generating the mesh based on the two-dimensional maps.
The patent also includes a non-transitory computer-readable medium encoded with codes that direct a processor to perform the steps outlined in the method. This medium can be used to implement the apparatus and method described in the patent.
Overall, this patent describes a system and method for generating a detailed three-dimensional mesh of a person using raw data and neural networks. The apparatus and method have potential applications in various fields, such as virtual reality, gaming, and animation, where accurate representation of human figures is important.