Daniel Lacko

Evaluation of an anthropometric shape model of the human scalp

This paper presents the evaluation a 3D shape model of the human head. A statistical shape model of the head is created from a set of 100 MRI scans. The ability of the shape model to predict new head shapes is evaluated by considering the prediction error distributions. The effect of using intuitive anthropometric measurements as parameters is examined and the sensitivity to measurement errors is determined. Using all anthropometric measurements, the average prediction error is 1.60 ± 0.36 mm, which shows the feasibility of the new parameters. The most sensitive measurement is the ear height, the least sensitive is the arc length. Finally, two applications of the anthropometric shape model are considered: the study of the male and female population and the design of a brain-computer interface headset. The results show that an anthropometric shape model can be a valuable tool for both research and design.

Correspondence Preserving Elastic Surface Registration with Shape Model Prior

In this paper, we describe a framework for surface registration. The framework consists of a combination of rigid registration, elasticity modulated registration and the use of a shape model prior. The main goal in this paper is to minimize the geometric surface registration error while maintaining correspondences. Experiments show improved geometric fit, correspondence, and timing compared to the current state of the art. Possible applications of the framework are construction of correspondences for shape models, reconstruction of missing parts, and artifact reduction.

Evaluation of 3D Body Shape Predictions Based on Features

The human body comes in many sizes and shapes. For design purposes, it is useful to be able to quickly simulate a virtual mannequin of a customer. A statistical shape model can be used for this purpose, because it describes the main variations of body shape inside the model’s population. From this model, the specific features of each person in the population are known. Therefore, a mapping between the shape model parameters and specific features can be calculated, which allows adjusting the body shape, in an intuitive way. In this work, we have investigated how accurate a body shape can be predicted based on a set of features and which features are most suitable for this purpose. Height, weight, and hip circumference appeared to be the most suitable features to accurately predict the body shape.

A Combined Statistical Shape Model of the Scalp and Skull of the Human Head

In this paper, we describe a framework to build a combined statistical shape model (SSM) of the outer surface of the scalp and the inner and outer surface of the skull of the human head. Such an SSM is a valuable tool when designing headgear, as it captures the variability of head geometry of a given population, enabling detailed analysis of the relation between the shape of the scalp and the skull. A combined SSM of the head may allow to work towards population based Finite Element (FE) models e.g. for safety and comfort predictions when wearing headgear. Therefore, a correspondence between the skull and scalp surfaces, originating from MRI scans, is determined using elastic surface registration. The combined SSM shown to be compact, to be able to generalize to unseen instances by adjusting the shape parameters and to be shape specific. Therefore, we can assure that, by adjusting the shape parameters, a broad range of realistic head shapes can be formed.

Thickness of Compressed Hair Layer: A Pilot Study in a Manikin

Recent advancements in 3D anthropometry enable to link a subject’s 1D measurements to its 3D shape to achieve better fit, functionality, comfort and/or safety. Statistical shape models of the human head retrieved from medical images (CT or MRI) allow for such parameterized models, with a recently proven accuracy of 1.6 mm with only four scalp parameters. This induces opportunities for better sizing systems and mass customization of head mounted products. Due to the nature of medical images, hair is never present in these models. Head shapes constructed from conventional optical 3D scans however do capture the presence of hair. In current optical 3D scans, subjects wear a swimming cap, hairnet or wig cap to flatten and compress the hair layer and prevent artifacts, interferences and misinterpretations during scanning due to the presence of hair. Thus models based on medical images provide parameterized scalp models with proven accuracy. This might be required for the design of certain head mounted wearables were sensors/actuators should make proper contact with the scalp through the hair. However, many other products to be designed for closely fitting the human head will rest on a (flattened) layer of hair. Thus in these situations it might be required to take account of the flattened and compressed hair layer in the design process. Up till now, knowledge of the compressed hair layer geometry is rarely needed in the field of industrial design with only anecdotic numbers available. The uprising of accurate parameter driven 3D models of the human head and entailed applications will induce the need for further accurate quantification thereof. Firstly, we present two a method to assess the thickness of the flattened and compressed hair layer by capturing hair thickness at a given point with a linear dial gauge. Secondly, we present another method to quantify the hair layer from points measured on scalp and hair layer. From this geometric information one can also deduce whether and to what extend the parametric scalp model approximates the head model with hair layer with the same accuracy as it approximates the scalp. Effect of hair thickness was evaluated by measuring a manikin with and without wigs and caps that flattened and compressed the wigs. All experiments were thus conducted in vitro, but both methods can be transferred without burdens to in vivo experiments. A hair thickness between 1.3 ± 0.4 mm was observed with the first method and 1.5 ± 1.3 mm with the second method. A small number of measurements indicate that the head form with flattened and compressed hair is predicted by the parametric scalp model with the same accuracy as it predicts the scalp form. These pilot results indicated that for the design of head mounted products that rest on the flattened hair layer, a margin of at least 1.5 mm should be taken into account for eventual variations in hair thickness. For the design of (personalized) head mounted products, already established parameterized scalp models can be used without loss of accuracy towards the presence of a flattened hair layer. Further large scale and in vivo studies are required to confirm and fine-tune these results.

Multi-patch B-Spline Statistical Shape Models for CAD-Compatible Digital Human Modeling

Parametric 3D human body models are valuable tools for ergonomic product design and statistical shape modelling (SSM) is a powerful technique to build realistic body models from a database of 3D scans. Like the underlying 3D scans, body models built from SSMs are typically represented with triangle meshes. Unfortunately, triangle meshes are not well supported by CAD software where spline geometry dominates. Therefore, we propose a methodology to convert databases of pre-corresponded triangle meshes into multi-patch B-spline SSMs. An evaluation on four 3D scan databases shows that our method is able to generate accurate and water-tight models while preserving inter-subject correspondences by construction. In addition, we demonstrate that such SSMs can be used to generate design manikins which can be readily used in SolidWorks for designing well conforming product parts.

Determining Comfortable Pressure Ranges for Wearable EEG Headsets

Measuring and interpretation of brain wave signals through electroencephalography (EEG) is an emerging technology. The technique is traditionally applied in a clinical setting with EEG caps and conductive gels to ensure proper contact through a subject’s hair, and anticipate inter-subject anthropometric variations. Development of dry electrodes offers the potential to develop wearable EEG headsets. Such devices could induce medical and commercial applications. In this paper, we evaluate a prototype EEG headset that actively places electrodes at standardized positions on the subject’s head, where each electrode is applied with equal pressure. The system is designed for use with dry electrodes. Our research delivers a better understanding on the link between general level of comfort and possible useful clear data signals, that can be used in brain computer interfaces (BCI). The present study is confined to the impact of adjustable electrodes pressure on level of user comfort only. Levels of discomfort are assessed in twelve participants, wearing an EEG headset with controllable electrode pressure exerted at 14 locations. Of-the-shelf dry electrodes are used. In a first session, evenly distributed pressure is increased and afterwards decreased in fixed time intervals, going from 10 kPa to 30 kPa and vice versa with steps of 2 kPa. In a second session, a subject specific acceptable pressure level is retrieved from the data of the first session and constantly applied for 30 min. During this intervention, level of discomfort is assessed in a VAS-scale. Additional observation and surveys yields insights on user experience in wearing a pressure exerting EEG headset.