Abstract:

Objective People are paying increasing attention to sleep quality, for which the scientific design of pillow shapes becomes more important. The shape of a pillow closely relates to the head, neck, and back of the human body. However, people of different ages have different body shapes, so existing pillows on the market hardly meet everyone's needs. This study focused on young adults and investigated the curved shape differences in their heads, necks, and backs.

Method The study adopted a point-line-surface-curvature classification approach. First, [TC]2 NX-16 non-contact 3-D scanner was used to capture human body point cloud data. Point cloud registration and iterative closest point (ICP) algorithm then extracted the head-neck-back median sagittal plane. Key points were marked according to surface morphological characteristics. After extracting horizontal and coronal planes based on feature point locations, a 3-D coordinate system was established to measure intersection coordinates. The study then performed data classification by integrating coordinate measurements with curvature parameters.

Results Three-dimensional scanning captured head-neck-back morphological data from 252 university male and female students aged 18-25. The posterior median sagittal plane revealed nine feature points, which are posterior occipital point, posterior neck point, axillary-sagittal intersection point, superior occipital peak point, intersection point of the cervical-occipital contour, inferior occipital peak point, and three cervicodorsal points. The spatial distribution of feature points guided the extraction of 11 curved surfaces (8 transverse, 3 longitudinal) across the head-neck-back regions. These surfaces formed 24 external contour intersection points in three-dimensional space.

A 3-D coordinate system was established with its origin at the center of the posterior neck point's horizontal cross-section, integrating spatial coordinates of all intersection points and median sagittal plane curvature parameters. Principal component analysis (PCA) revealed that the first four principal components achieved a cumulative variance contribution rate of 0.8, effectively reducing data dimensionality while preserving most information. The PCA algorithm identified four characteristic factors. K-means clustering analysis sequentially calculated sum of squared errors values between each sample and its cluster center for different K values. Optimal clustering occurred at K=4, leading to the final classification of head-neck-back surface morphology into four characteristic types, which are rounded, deep-necked, flattened, and forward-leaning.

The results showed that the occipital bone line of the round type was smooth and natural with a uniform transition. The neck depression of the deep-necked type was relatively deep, forming an obvious drop compared with the surrounding areas. The curvature change of the occipital bone of the flattened type was small and almost flat. The entire head-neck-back of the forward-leaning type demonstrated a forward leaning trend, deviating from the normal physiological curve. According to the morphological discrimination rules, they re-determined the initial sample data as a whole, and the accuracy rate reached as high as 95.58%, showing a remarkable effect.

Conclusion The study utilized the median sagittal plane as the anatomical reference for head-neck-back morphological analysis, establishing a three-dimensional Cartesian coordinate system. Principal component analysis combined with K-means clustering of spatial coordinates and curvature parameters classified the samples into four distinct morphological types: rounded, deep-necked, flattened, and forward-leaning, demonstrating a classification accuracy of 95.58%. The curved pillow design incorporated personalized parameters for different morphological types. The rounded type was characterized by its smooth and curvilinear contour, and an increased groove depth is essential. This enhancement allows the pillow to cradle the head's curvature snugly during sleep, reducing pressure points and promoting a more relaxed rest. The deep-necked type with pronounced cervical concavity benefits significantly from an elevated pillow apex height. This adjustment ensures that the pillow aligns perfectly with the natural curve of the neck, providing crucial support and preventing strain. The flattened type presents a relatively even surface, an expanded contact surface area is imperative. This design feature distributes pressure evenly across the occipital region, minimizing discomfort and enhancing sleep quality. For the forward-leaning type, a raised front section with a gradual slope is incorporated into the pillow design. This unique structure gently corrects spinal alignment during rest, helping to restore the natural posture and alleviate potential back pain.

Key words: head-neck-back curve, ergonomics, three-dimensional human body scanning, morphological recognition, curved-pillow, personalized customization