Novel 4D Whole Body Scanning Solution and its Medical Application

Abstract

One of the drawbacks of the traditional 3D whole body scanning is that it is capable of capturing only static models. In most of the cases it is impossible to properly analyze the way people move as they are not able to freeze their movement for a certain amount of time. In order to add a 4th dimension (time) to measurements, a system have to be built using stable ultra-fast 3D scanners. The presented solution meets industrial requirements for 4D measurements of dynamic objects. It is capable of acquiring up to 120 Hz sequences of high precision point clouds along with an information about its lightness and normal vectors. A spatial resolution of 1 mm is obtained with an inaccuracy below 0.5 mm. It was originally designed for a 4D human body shape measurement to support medical rehabilitation monitoring, however it is not restricted to this application. The system is composed of four directional measurement columns [1]. Sufficient body surface coverage is possible thanks to an even distribution of modules, each consisting of 1 projector and 2 detectors-on the upper and lower part of the head. Their working principle is based on a structured light projection, specifically a single frame pattern approach which enabled achievement of the declared frequency. For this particular case a problem of synchronization (highly erroneous overlapping of the projected patterns) was solved. A sine modulated patterns are colored and distinguished using a spectral separation via color filters. Information about fringe numbers is encoded using an additional transverse modulation of the patterns. Retrieval of a single multidirectional output cloud is done using a set of dedicated algorithms, including phase unwrapping on a single image per detector, scaling into XYZ coordinates and common calibration. The high precision 4D data is very heavy. A raw 1 minute of 120 Hz scan requires around 360 GB of a disk space. In order to handle such data, the specialized software called FRAMES (Framework and Robust Algorithms for Models of Extreme Size) was developed. It has built-in 4D RAM (Random Access Memory) manager which enables efficient visualization, advanced multithread processing and analysis of such data. The presented 4D scanning solution was tested in a real-life environment. The possibility of performing 3D body scanning in time enabled the rehabilitation progress monitoring after leg amputation.

Decreased Vertical Trunk Inclination Angle and Pelvic Inclination as the Result of Mid-High-Heeled Footwear on Static Posture Parameters in Asymptomatic Young Adult Women

“…Dedicated algorithms were developed to detect on-skin fiducial markers and calculate the parameters for evaluating changes in posture and balance according to current recommendations [42,60,61,62]. Analysis of the measurement data was performed using IMP’s proprietary FRAMES programming environment [63], an extensible framework used for different tasks involving point cloud data. It was written in C++ with minimal use of third-party programming libraries.…”

sSfS: Segmented Shape from Silhouette Reconstruction of the Human Body

“…Reconstruction accuracy, quality, and speed have improved significantly over recent years to the point where four-dimensional (4D, i.e., 3D over time) dynamic movement scanners (e.g., Diers International GmbH [ 1 , 2 ], 3dMD LLC [ 3 , 4 ], and Microsoft Corporation [ 5 ]) have become an industry standard. Both dynamic and static whole-body scanners have countless applications, such as in special effects in movies/video games [ 6 ], in human body pose and deformation analyses for medical diagnostics [ 7 , 8 , 9 ], and in anthropometry [ 10 ]. However, most of these systems have various limitations, i.e., insufficient precision, high cost, and noisy or incomplete data.…”