Hand Configuration Research Articles

Effects of Handedness and Viewpoint on the Imitation of Origami-Making

The evolutionary origins of the human bias for 85% right-handedness are obscure. The Apprenticeship Complexity Theory states that the increasing difficulty of acquiring stone tool-making and other manual skills in the Pleistocene favoured learners whose hand preference matched that of their teachers. Furthermore, learning from a viewing position opposite, rather than beside, the demonstrator might be harder because it requires more mental transformation. We varied handedness and viewpoint in a bimanual learning task. Thirty-two participants reproduced folding asymmetric origami figures as demonstrated by a videotaped teacher in four conditions (left-handed teacher opposite the learner, left-handed beside, right-handed opposite, or right-handed beside). Learning performance was measured by time to complete each figure, number of video pauses and rewinds, and similarity of copies to the target shape. There was no effect of handedness or viewpoint on imitation learning. However, participants preferred to produce figures with the same asymmetry as demonstrated, indicating they imitate the teacher's hand preference. We speculate that learning by imitation involves internalising motor representations and that, to facilitate learning by imitation, many motor actions can be flexibly executed using the demonstrated hand configuration. We conclude that matching hand preferences evolved due to socially learning moderately complex bimanual skills.

Cortical control of object-specific grasp relies on adjustments of both activity and effective connectivity: a common marmoset study.

Key points The cortical mechanisms of grasping have been extensively studied in macaques and humans; here, we investigated whether common marmosets could rely on similar mechanisms despite strong differences in hand morphology and grip diversity.We recorded electrocorticographic activity over the sensorimotor cortex of two common marmosets during the execution of different grip types, which allowed us to study cortical activity (power spectrum) and physiologically inferred connectivity (phase‐slope index).Analyses were performed in beta (16–35 Hz) and gamma (75–100 Hz) frequency bands and our results showed that beta power varied depending on grip type, whereas gamma power displayed clear epoch‐related modulation.Strength and direction of inter‐area connectivity varied depending on grip type and epoch.These findings suggest that fundamental control mechanisms are conserved across primates and, in future research, marmosets could represent an adequate model to investigate primate brain mechanisms. The cortical mechanisms of grasping have been extensively studied in macaques and humans. Here, we investigated whether common marmosets could rely on similar mechanisms despite striking differences in manual dexterity. Two common marmosets were trained to grasp‐and‐pull three objects eliciting different hand configurations: whole‐hand, finger and scissor grips. The animals were then chronically implanted with 64‐channel electrocorticogram arrays positioned over the left premotor, primary motor and somatosensory cortex. Power spectra, reflecting predominantly cortical activity, and phase‐slope index, reflecting the direction of information flux, were studied in beta (16–35 Hz) and gamma (75–100 Hz) bands. Differences related to grip type, epoch (reach, grasp) and cortical area were statistically assessed. Results showed that whole‐hand and scissor grips triggered stronger beta desynchronization than finger grip. Task epochs clearly modulated gamma power, especially for finger and scissor grips. Considering effective connectivity, finger and scissor grips evoked stronger outflow from primary motor to premotor cortex, whereas whole‐hand grip displayed the opposite pattern. These findings suggest that fundamental control mechanisms, relying on adjustments of cortical activity and connectivity, are conserved across primates. Consistently, marmosets could represent a good model to investigate primate brain mechanisms.

Passive and Space-Discriminative Ionic Sensors Based on Durable Nanocomposite Electrodes toward Sign Language Recognition.

This work developed an ionic sensor for human motion monitoring by employing durable H-reduced graphene oxide (RGO)/carbon nanotubes (CNTs)/Ag electrodes and an ionic polymer interlayer. The sensor functions as a result of unbalanced ion transport and accumulation between two electrodes stimulated by applied deformation. The networking structure and stable electrodes provide convenient ion-transport channels and a large ion accumulation space, resulting in a sensitivity of 2.6 mV in the strain range below 1% and high stability over 6000 bending cycles. Ionic sensors are of intense interest motivated by detecting human activities, which usually associate with a large strain or deformation change. More importantly, direction identification and spatial deformation recognition are feasible in this research, which is beneficial for the detection of complex multidimensional activities. Here, an integrated smart glove with several sensors mounted on the hand joints displays a distinguished ability in the complex geometry of hand configurations. Based on its superior performance, the potential applications of this passive ionic sensor in sign language recognition and human-computer interaction are demonstrated.

Conceptualizing the phonology of sign language: As a matter of emphasis and elucidation

Like any other language, sign language is, among other things, guided by phonological principles involving several parameters combining in space to form lexical signs. According to Klima and Bellugi (1979), a simple lexical sign is essentially a simultaneous occurrence of a particular value of each of the several parameters. These parameters occur within a constrained signing space in combination with each other. William Stokoe (1960) equated phonology in oral language to what he termed cherology in sign language. However, most sign linguists prefer to use the term phonology. In oral language, phonology is the study of sounds, but in sign language, it is concerned with the parameters that make sign language a systematic and intelligible language system. One basic difference between oral and sign language is that in oral language we use words while in sign language we use signs. The other important difference is that while oral language is produced in the oral cavity using articulatory organs such as the mouth, sign language is produced in space using mostly the hands. Yet the structural difference is that while a phoneme in oral language is the basic unit of a word a chereme in sign language is the basic unit of a sign. In oral language words are organized sequentially while the signs in sign language are organized as a combination of simultaneously occurring components derived from several spatial dimensions. The purpose of this paper is to examine the parameters that characterize sign language and to demonstrate the formation, vocabulary and organization of signs. The paper demonstrates how signs are executed in space. The paper addresses four major phonological (cherological) parameters of sign language, namely hand configuration (hand shape), place of articulation (location / position), movement and orientation. Major parameters save to distinguish very large classes of signs. In addition the paper articulates how these parameters combine to formulate meaningful signs within the constrained signing space. From the analysis, the author concludes that the phonology of sign language is highly complex and is resident in space as a function of the hand acting as a highly articulatory linguistic organ.

Os dicionários da língua brasileira de sinais e suas contribuições

Os dicionários de línguas de sinais atualmente são inúmeros, sendo impressos ou em formatos digitais, fabricados por surdos, instituições de ensino que capacitam profissionais na língua de sinais ou por iniciativas privadas. É uma realidade necessária diante da complexidade dos sinais e, além disso, é também, através deles que se torna possível manter a padronização necessária durante a conversação ou atuação dos profissionais tradutores intérpretes de língua de sinais. Esses dicionários, sejam impressos ou digitais, descrevem informações fonológicas, gramaticais e semânticas acerca dos sinais e das palavras, que de fato facilitam e permitem melhor compreensão do sinal pesquisado. Nos dicionários impressos é possível perceber variações quanto às formas encontradas, tais como: foto, descrição dos sinais, escrita de sinais, ilustrações e tradução para a língua oral. Já nos dicionários digitais, há possibilidade de busca por ordem alfabética do português ou pela configuração de mão, no qual os sinais são representados por filmagens, contendo assim sua descrição e definição. Este artigo traz um histórico dos dicionários registrados na língua de sinais, relatando a influência da França no primeiro dicionário de Libras no Brasil, incita reflexões sobre a relevância dos dicionários existentes e observa a necessidade da criação de dicionários nas áreas de especialidades, visto a considerável ausência de sinais para diversos termos usados no cotidiano acadêmico.

Cooperative gestures for industry: Exploring the efficacy of robot hand configurations in expression of instructional gestures for human–robot interaction

Fast and reliable communication between human worker(s) and robotic assistants is essential for successful collaboration between the agents. This is especially true for typically noisy manufacturing environments that render verbal communication less effective. In this work, we investigate the efficacy of nonverbal communication capabilities of robotic manipulators that have poseable, three-fingered end-effectors (hands). We explore the extent to which different poses of a typical robotic gripper can effectively communicate instructional messages during human–robot collaboration. Within the context of a collaborative car door assembly task, we conducted a series of three studies. We first observed the type of hand configurations that humans use to nonverbally instruct another person (Study 1, N = 17); based on the observation, we examined how well human gestures with frequently used hand configurations are understood by recipients of the message (Study 2, N = 140). Finally, we implemented the most human-recognized human hand configurations on a seven-degree-of-freedom robotic manipulator to investigate the efficacy of having human-inspired hand poses on a robotic hand compared to an unposed hand (Study 3, N = 100). Contributions of this work include presentation of a set of hand configurations humans commonly use to instruct another person in a collaborative assembly scenario, as well as recognition rate and recognition confidence measures for the gestures that humans and robots express using different hand configurations. Results indicate that most gestures are better recognized with a higher level of confidence when displayed with a posed robot hand.

Concurrent anticipation of two object dimensions during grasping in 10-month-old infants: A quantitative analysis

The anticipation of more than one object dimension while grasping for objects has been rarely investigated in infancy. The few existing studies by Newell et al. and Schum et al. have revealed mixed results probably mainly due to methodological limitations. Therefore, the present experiments tested concurrent anticipatory grasping for two object dimensions, namely, object size and object orientation using a quantitative motion capture system (Vicon), in 10-month-old infants and adults. We presented objects varying in size (small vs. large) and orientation (horizontally vs. vertically) and analyzed participants’ anticipatory hand configurations. As with adults, we observed that infants rotated their wrists, thumbs, and index fingers as a function of object orientation and adjusted their maximum grip apertures and their grip apertures shortly before they touched the objects as a function of object size. Analyses on an individual level showed that infants like adults anticipated both dimensions when the maximal values of aperture and angle were used but not when the measures shortly before touch were considered. Thus, the ability to anticipate more than one object dimension can already be observed at 10 months of age but seems to improve considerably over the first year of life.

Relation of anthropometric hand measurements to idiopathic carpal tunnel syndrome

ContextCertain individuals are more prone to developing idiopathic carpal tunnel syndrome (ICTS) than others, suggesting that certain personal factors can be implicated in its occurrence.AimThe aim of this work was to study anthropometric hand and wrist measurements in ICTS patients, and to correlate them with median nerve electrophysiologic study.Patients and methodsThe study included 50 patients with clinically diagnosed and electrophysiologically confirmed ICTS and 50 age-matched and sex-matched healthy volunteers as the control group. Both groups performed sensory and motor conduction studies of the median nerve. External hand and wrist anthropometric measurements were taken for both groups including wrist depth and width, wrist ratio (WR), palm length and width, third digit length, and hand ratio (HR).ResultsPatients had significantly higher wrist depth (P=0.000), higher WR (P=0.000), shorter palm length (P=0.002), shorter hand length (P=0.001), and lower HR (P=0.000). Patients had more square wrists and shorter hands. Some of these measurements correlated well with median nerve conduction study parameters. Wrist depth and WR were positively correlated with median motor and sensory latencies (P=0.000) and negatively correlated with median motor and sensory amplitudes (P=0.000), and sensory conduction velocity (P=0.000). HR was negatively correlated to median motor and sensory latencies (P=0.01) and positively correlated to median motor (P=0.001) and sensory amplitudes (P=0.039), and sensory conduction velocity (P=0.001). Palm width was negatively correlated with median motor amplitude (P=0.043).ConclusionCertain hand anthropometric characteristics predispose to ICTS. Short hand and square wrist configurations could predict the development of ICTS.

A fully automatic method for recognizing hand configurations of Brazilian sign language

Abstract Introduction Sign language is a collection of gestures, postures, movements, and facial expressions used by deaf people. The Brazilian sign language is Libras. The use of Libras has been increased among the deaf communities, but is still not disseminated outside this community. Sign language recognition is a field of research, which intends to help the deaf community communication with non-hearing-impaired people. In this context, this paper describes a new method for recognizing hand configurations of Libras - using depth maps obtained with a Kinect® sensor. Methods The proposed method comprises three phases: hand segmentation, feature extraction, and classification. The segmentation phase is independent from the background and depends only on pixel value. The feature extraction process is independent from rotation and translation. The features are extracted employing two techniques: (2D)2LDA and (2D)2PCA. The classification employs two classifiers: a novelty classifier and a KNN classifier. A robust database is constructed for classifier evaluation, with 12,200 images of Libras and 200 gestures of each hand configuration. Results The best accuracy obtained was 96.31%. Conclusion The best gesture recognition accuracy obtained is much higher than the studies previously published. It must be emphasized that this recognition rate is obtained for different conditions of hand rotation and proximity of the depth camera, and with a depth camera resolution of only 640×480 pixels. This performance must be also credited to the feature extraction technique, and to the size standardization and normalization processes used previously to feature extraction step.

Decoding Grasping Movements from the Parieto-Frontal Reaching Circuit in the Nonhuman Primate

Prehension movements typically include a reaching phase, guiding the hand toward the object, and a grip phase, shaping the hand around it. The dominant view posits that these components rely upon largely independent parieto-frontal circuits: a dorso-medial circuit involved in reaching and a dorso-lateral circuit involved in grasping. However, mounting evidence suggests a more complex arrangement, with dorso-medial areas contributing to both reaching and grasping. To investigate the role of the dorso-medial reaching circuit in grasping, we trained monkeys to reach-and-grasp different objects in the dark and determined if hand configurations could be decoded from functional magnetic resonance imaging (MRI) responses obtained from the reaching and grasping circuits. Indicative of their established role in grasping, object-specific grasp decoding was found in anterior intraparietal (AIP) area, inferior parietal lobule area PFG and ventral premotor region F5 of the lateral grasping circuit, and primary motor cortex. Importantly, the medial reaching circuit also conveyed robust grasp-specific information, as evidenced by significant decoding in parietal reach regions (particular V6A) and dorsal premotor region F2. These data support the proposed role of dorso-medial "reach" regions in controlling aspects of grasping and demonstrate the value of complementing univariate with more sensitive multivariate analyses of functional MRI (fMRI) data in uncovering information coding in the brain.

Distinct Contributions of Dorsal and Ventral Streams to Imitation of Tool-Use and Communicative Gestures.

Imitation of tool-use gestures (transitive; e.g., hammering) and communicative emblems (intransitive; e.g., waving goodbye) is frequently impaired after left-hemispheric lesions. We aimed 1) to identify lesions related to deficient transitive or intransitive gestures, 2) to delineate regions associated with distinct error types (e.g., hand configuration, kinematics), and 3) to compare imitation to previous data on pantomimed and actual tool use. Of note, 156 patients (64.3 ± 14.6 years; 56 female) with first-ever left-hemispheric ischemic stroke were prospectively examined 4.8 ± 2.0 days after symptom onset. Lesions were delineated on magnetic resonance imaging scans for voxel-based lesion-symptom mapping. First, while inferior-parietal lesions affected both gesture types, specific associations emerged between intransitive gesture deficits and anterior temporal damage and between transitive gesture deficits and premotor and occipito-parietal lesions. Second, impaired hand configurations were related to anterior intraparietal damage, hand/wrist-orientation errors to premotor lesions, and kinematic errors to inferior-parietal/occipito-temporal lesions. Third, premotor lesions impacted more on transitive imitation compared with actual tool use, pantomimed and actual tool use were more susceptible to lesioned insular cortex and subjacent white matter. In summary, transitive and intransitive gestures differentially rely on ventro-dorsal and ventral streams due to higher demands on temporo-spatial processing (transitive) or stronger reliance on semantic information (intransitive), respectively.

Evaluation and Optimal Selection of Bionic Hand Configurations

Evaluation and Optimal Selection of Bionic Hand Configurations

New Tools of Alternative Communication for Persons with Verbal Communication Disorders

This paper presents an information technology based on modern IT devices (netbooks, tablets, phones, etc) and supporting an alternative communication for persons with (temporary or permanent) verbal communication disorders. It is proposed to implement this communication through the intellectualization of the process of entering text data with applying a limited number of relevant controls specified by a person's hand. A human skeleton model is investigated with a view to determining a limited number of hand configurations and motions reliably identified by recognition tools of IT devices.

EPV 19. Dissociable regions for recognition and execution of conceptual and spatio-temporal action characteristics in acute stroke patients

The ability to recognize and respond to actions performed by others is fundamental for a wide range of activities involving social interactions. However, despite an enormous number of functional imaging studies, a precise delineation of the brain regions crucial for distinct aspects of action recognition has remained elusive. Moreover, the degree of overlap between regions involved in active action performance is controversial. To elucidate these issues, 98 acute stroke patients (68 male, age mean ± SD, 65 ± 13 years) were examined in the acute period after stroke (mean ± SD 4.4 ± 2 days after symptom onset). First, action recognition was assessed with a newly designed test which required patients to distinguish between correctly and incorrectly performed tool-associated actions depicted in short videos. Errors categories comprised conceptual (e.g., spreading jam on toast with a paint brush), and spatio-temporal errors (orientation of the tool, hand configuration, and movement kinematics). Second, actual use of tools was probed using an established test which, in analogy to the error categories for action recognition, included separate measures for the ability to match tools and recipients (e.g., hammer – nail; conceptual score), and for the quality of the executed action (e.g., hammering; spatio-temporal score). Lesions were delineated on diffusion-weighted scans for voxel-based lesion-symptom mapping. Impaired recognition of conceptual errors was associated mainly with lesions to the anterior middle temporal gyrus as well as to the caudal middle frontal gyrus. Conversely, the detection of spatio-temporal errors depended on the posterior superior temporal gyrus and sulcus, inferior parietal lobule (IPL) and ventral premotor cortex (vPMC). Analyses for distinct spatio-temporal error subcategories revealed that superior temporal lesions specifically affected the recognition of kinematic and tool orientation errors, while inferior parietal and premotor damage selectively impaired the discrimination of hand configurations. Lastly, compared to the recognition of conceptual and spatio-temporal errors, both active tool-recipient matching and action execution, respectively, were more dependent on the IPL and vPMC; conversely, action recognition deficits were more strongly associated with occipito-temporal and superior temporal lesions. The results demonstrate that action recognition is enabled by a network of dissociable regions that are specialized for the processing of distinct action features. While the decoding of conceptual action aspects relies mainly on anterior temporal regions, IPL and vPMC mediate the recognition of hand configurations, and the posterior superior temporal lobe is crucial for the processing of movement kinematics and tool orientation. Our findings show that, with the exception of the hand configuration, action recognition does not rely on the fronto-parietal regions that subserve the performance of actual tool use.

O desenvolvimento da percepção dos contrastes mínimos na língua brasileira de sinais em um grupo de Codas

RESUMO Objetivo: verificar o surgimento e a estabilização da percepção dos contrastes mínimos da Língua Brasileira de Sinais, crianças ouvintes filhas de pais surdos. Métodos: foram analisadas nove coletas das habilidades perceptuais em Língua Brasileira de Sinais de crianças ouvintes, filhas de pais surdos com idades entre dois e nove anos. O instrumento aplicado foi composto de 35 pares mínimos, sinais que variavam em somente um parâmetro, podendo esse ser: configuração de mão, locação, movimento ou orientação. Realizou-se análise estatística com nível de significância de 5%, e foram utilizados o teste de comparação de Friedman e Wilcoxon, além da correlação de Spearman. Resultados: o parâmetro movimento é percebido mais facilmente do que os demais contrastes. Seguido dos parâmetros locação e configuração de mão, que atuam de maneira semelhante na percepção dos aprendizes desta língua. O contraste mais difícil de ser percebido refere-se à orientação. Quanto mais tempo de contato com a língua, melhor o desempenho das Codas na percepção dos contrastes mínimos. Conclusão: por meio do instrumento de percepção constatou-se quais os parâmetros são percebidos primeiramente pelos aprendizes. O parâmetro movimento foi mais facilmente percebido, seguido da locação e configuração de mão, já a orientação foi o último a ser adquirido.

A new method for recognizing hand configurations of Brazilian gesture language.

This paper describes a new method for recognizing hand configurations of the Brazilian Gesture Language - LIBRAS - using depth maps obtained with a Kinect® camera. The proposed method comprised three phases: hand segmentation, feature extraction, and classification. The segmentation phase is independent from the background and depends only on pixel depth information. Using geometric operations and numerical normalization, the feature extraction process was done independent from rotation and translation. The features are extracted employing two techniques: (2D)2LDA and (2D)2PCA. The classification is made with a novelty classifier. A robust database was constructed for classifier evaluation, with 12,200 images of LIBRAS and 200 gestures of each hand configuration. The best accuracy obtained was 95.41%, which was greater than previous values obtained in the literature.

Disentangling Representations of Object and Grasp Properties in the Human Brain

The properties of objects, such as shape, influence the way we grasp them. To quantify the role of different brain regions during grasping, it is necessary to disentangle the processing of visual dimensions related to object properties from the motor aspects related to the specific hand configuration. We orthogonally varied object properties (shape, size, and elongation) and task (passive viewing, precision grip with two or five digits, or coarse grip with five digits) and used representational similarity analysis of functional magnetic resonance imaging data to infer the representation of object properties and hand configuration in the human brain. We found that object elongation is the most strongly represented object feature during grasping and is coded preferentially in the primary visual cortex as well as the anterior and posterior superior-parieto-occipital cortex. By contrast, primary somatosensory, motor, and ventral premotor cortices coded preferentially the number of digits while ventral-stream and dorsal-stream regions coded a mix of visual and motor dimensions. The representation of object features varied with task modality, as object elongation was less relevant during passive viewing than grasping. To summarize, this study shows that elongation is a particularly relevant property of the object to grasp, which along with the number of digits used, is represented within both ventral-stream and parietal regions, suggesting that communication between the two streams about these specific visual and motor dimensions might be relevant to the execution of efficient grasping actions. To grasp something, the visual properties of an object guide preshaping of the hand into the appropriate configuration. Different grips can be used, and different objects require different hand configurations. However, in natural actions, grip and object type are often confounded, and the few experiments that have attempted to separate them have produced conflicting results. As such, it is unclear how visual and motor properties are represented across brain regions during grasping. Here we orthogonally manipulated object properties and grip, and revealed the visual dimension (object elongation) and the motor dimension (number of digits) that are more strongly coded in ventral and dorsal streams. These results suggest that both streams play a role in the visuomotor coding essential for grasping.

Dexterous Grasping Tasks Generated With an Add-on End Effector of a Haptic Feedback System

The simulation of grasping operations in virtual reality (VR) is required for many applications, especially in the domain of industrial product design, but it is very difficult to achieve without any haptic feedback. Force feedback on the fingers can be provided by a hand exoskeleton, but such a device is very complex, invasive, and costly. In this paper, we present a new device, called HaptiHand, which provides position and force input as well as haptic output for four fingers in a noninvasive way, and is mounted on a standard force-feedback arm. The device incorporates four independent modules, one for each finger, inside an ergonomic shape, allowing the user to generate a wide range of virtual hand configurations to grasp naturally an object. It is also possible to reconfigure the virtual finger positions when holding an object. The paper explains how the device is used to control a virtual hand in order to perform dexterous grasping operations. The structure of the HaptiHand is described through the major technical solutions required and tests of key functions serve as validation process for some key requirements. Also, an effective grasping task illustrates some capabilities of the HaptiHand.

Design and Evaluation of Visual Interpenetration Cues in Virtual Grasping

We present design and impact studies of visual feedback for virtual grasping. The studies suggest new or updated guidelines for feedback. Recent grasping techniques incorporate visual cues to help resolve undesirable visual or performance artifacts encountered after real fingers enter a virtual object. Prior guidelines about such visuals are based largely on other interaction types and provide inconsistent and potentially-misleading information when applied to grasping. We address this with a two-stage study. In the first stage, users adjusted parameters of various feedback types, including some novel aspects, to identify promising settings and to give insight into preferences regarding the parameters. In the next stage, the tuned feedback techniques were evaluated in terms of objective performance (finger penetration, release time, and precision) and subjective rankings (visual quality, perceived behavior impact, and overall preference). Additionally, subjects commented on the techniques while reviewing them in a final session. Performance wise, the most promising techniques directly reveal penetrating hand configuration in some way. Subjectively, subjects appreciated visual cues about interpenetration or grasp force, and color changes are most promising. The results enable selection of the best cues based on understanding the relevant tradeoffs and reasonable parameter values. The results also provide a needed basis for more focused studies of specific visual cues and for choosing conditions in comparisons to other feedback modes, such as haptic, audio, or multimodal. Considering results, we propose that 3D interaction guidelines must be updated to capture the importance of interpenetration cues, possible performance benefits of direct representations, and tradeoffs involved in cue selection.

Improving Kinematic Accuracy of Soft Wearable Data Gloves by Optimizing Sensor Locations.

Bending sensors enable compact, wearable designs when used for measuring hand configurations in data gloves. While existing data gloves can accurately measure angular displacement of the finger and distal thumb joints, accurate measurement of thumb carpometacarpal (CMC) joint movements remains challenging due to crosstalk between the multi-sensor outputs required to measure the degrees of freedom (DOF). To properly measure CMC-joint configurations, sensor locations that minimize sensor crosstalk must be identified. This paper presents a novel approach to identifying optimal sensor locations. Three-dimensional hand surface data from ten subjects was collected in multiple thumb postures with varied CMC-joint flexion and abduction angles. For each posture, scanned CMC-joint contours were used to estimate CMC-joint flexion and abduction angles by varying the positions and orientations of two bending sensors. Optimal sensor locations were estimated by the least squares method, which minimized the difference between the true CMC-joint angles and the joint angle estimates. Finally, the resultant optimal sensor locations were experimentally validated. Placing sensors at the optimal locations, CMC-joint angle measurement accuracies improved (flexion, 2.8° ± 1.9°; abduction, 1.9° ± 1.2°). The proposed method for improving the accuracy of the sensing system can be extended to other types of soft wearable measurement devices.