Healthcare. Where we began. Imaging.

Advancing medical and dental innovation incorporates 3dMD’s most mature customer community. In 2000, 3dMD assisted in transitioning the health profession from taking traditional 2D photographs as a visual reference to capturing anatomically-precise 3D surface images (non-invasively). 3dMD images support the clinician when he/she is assessing surgical intervention, objectively measuring morphological change throughout the treatment cycle, and evaluating outcomes. From its early origins, 3dMD had the task of delivering technology products that work seamlessly in challenging pediatric hospitals to assist with complex long-term treatment cycles such as cleft lip/palate. Today with 3dMD’s dynamic-3D/4D motion systems, 3dMD customer teams can assess, quantify, and measure shape change in a patient’s anatomical function, movement, pose, posture, communication, and expression. In medical as well as the dental worlds, 3dMD systems are now considered the standard ‘near-ground truth’ 3D image reference system for 3D anthropometric analysis. To date, 3dMD has been used and referenced supporting hundreds of articles in peer-reviewed scientific journals, as well as presentations and posters at medical and dental conferences all around the world.

Computer Vision / Perception. Training.

With a challenging objective of automating computer programs to replicate the responsibilities of a human visual system, Computer Vision focuses on developing a deep machine learning model. Today advanced 3dMD customer teams are focused on collecting and building databases of ‘near-ground truth’ 3D shape image information (Real-World) for synthesis into a generic factual training phase to ensure the foundation of best output results. As a baseline, 3dMD is essentially a computer vision application. However, 3dMD is also a tool to help research teams rapidly advance their own computer vision applications in the early phase of development. With a medical imaging foundation, 3dMD systems provide teams with a progressive sequence of Real-World 3dMD images whose provenance can be related to a predictable and consistent ‘near-ground truth’ derivative. This will enable research teams to focus on establishing deep learning resources that document quantitatively observed behavior for perceptive interpretation in the wild –  whether it is to achieve contactless explicit communication with devices or teach devices to interpret and predict implicit human intent. 3dMD provides customer teams not only with Real-World data of people, more importantly, Real-World data of the human anatomy of subjects in motion with ‘near-ground-truth’ 3D shape information for all 3dMD images in the progressive sequence.

Artificial Intelligence | Machine Learning. Training.

Human visual perception inputs sensory experiences into a heuristic analysis to make qualitive decisions. To support AI and ML initiatives, 3dMD customer teams are building vast diverse Real-World databases of human subjects performing natural actions; gesturing and communicating with their face, body, limbs, and/or hands; and reacting to stimuli to build the deep machine learning model needed to quantitatively characterize visual sensory experiences. 3dMD systems are both quantitative and qualitative, feeding the deep learning model rich, Real-World ‘near-ground truth’ 3D shape information captured at 10-120Hz throughout the entire subject performance. Because of the depth and richness of the Real-World data collected it has a persistent value and can be re-mined to explore further detail, flow, and structure as the AI Kernel develops. 3dMD provides customer teams not only with Real-World data of people, more importantly, Real-World data of the human anatomy of subjects in motion with ‘near-ground-truth’ 3D shape information for all 3dMD images in the progressive sequence.

Humanoid Robotics. Training.

Training Humanoids with Real Human Movement — Hands, Face, and Full Body Precision.
3dMD, a proven standard in supporting patient care and clinical research in the health sector, delivers the world’s most accurate dynamic-3D/4D human imaging technology… making it an obvious fundamental data source for training next-generation humanoid robots.

Most attention in the humanoid robotics space has focused on manufacturer hardware: joints, actuators, and walking gaits. But, just like people, the real intelligence will live in the humanoid software brain platform… advanced control systems that will guide perception, interaction, and movement. To train the humanoid brain, 3dMD provides its customer teams with dynamic, medical-grade 3D/4D imaging systems for capturing performance and variance in human bodies, hands, and faces in natural motion and with anatomical accuracy. 3dMD real-world data helps reduce AI training errors, fill critical gaps in synthetic datasets, and close the sim-to-real divide.

Today’s humanoid robotics platforms are evolving rapidly to incorporate human-level dexterity, perception, and interaction. To achieve this fidelity and comprehension, humanoid robots must be trained on much more than 2D videos and synthetic approximations of human motion. They need to assimilate actual real-world human behavior… captured in its full 3D/4D complexity. Enter 3dMD.

3dMD Accelerating Humanoid Robotics
High-Resolution 360-degree Body Data. Capture dynamic-3D/4D image sequences of subjects with sub-millimeter precision, which is ideal for training natural locomotion, balance, and spatial reasoning systems, as well as recognizing gesture, gait, and body language.

Detailed Hand Data for Fine-Motor Skill Learning. Capture dynamic-3D/4D image sequences of real hands performing including grasping and releasing objects, using precision tools and delicate interaction, and coordinating bimanual actions. Unlike traditional motion-capture, videos, or synthetic datasets, 3dMD delivers anatomical 3D/4D ground-truth surface geometry, which is essential for robots learning tactile and dexterous hand control.

High-Fidelity Face Data for Human-Robot Interaction Learning. Capture dynamic-3D/4D image sequences of facial expressions, functions, and speech including subtle muscle movements, dialogue, eye gaze, and social gestures. These 3dMD sequences are critical for teaching humanoid robots to recognize and respond to human emotion, maintain socially appropriate engagement, and mirror or model facial expressions and cues. Whether your humanoid robot needs to work in customer-facing roles or assist in home environments, face-level intelligence starts with accurate, expressive training data.

Imitation Learning from Real Digitized Human Behavior. Enable humanoid robots to learn from demonstration using complete human task 3D/4D 3dMD image sequences… all captured across diverse people and behaviors.

Real-World 3dMD Data Advantage
• Trusted medical-grade ground truth 3D-shape accuracy and visual fidelity: used in pediatric treatment, surgery, biomechanics, and research
• True 3D surface geometry with no markers, suits, or synthetic rigs
• Diverse real-world 3D/4D datasets across age, ethnic background, gender, body types, and motion styles
• Optimized for feeding demanding AI pipelines for robotics, machine learning, AR/VR, and digital human modeling

Multi-Purpose 3dMD Enrollment of Human Subjects
• Train dexterous hands with real human grip dynamics
• Teach general-purpose robotic devices to read and respond to facial expressions
• Generate task demonstrations for imitation learning
• Improve vision and perception AI input for real-world interaction
• Reduce sim-to-real error with 3D/4D ground-truth, natural human movement
• Optimize ergonomics of wearable robotic devices
• Seamlessly integrate soft robotics into co-existent virtual environments

Powering the Next Generation of Humanoid Intelligence
Whether your team is building humanoid robots for manufacturing, logistics, personal assistance, or general-purpose cognition, 3dMD offers a robust, real-world data foundation for training perception, movement, and behavior. Real-world 3dMD data goes beyond teaching humanoids how to just move, it captures rich ground truth data teaching subtleties and nuances associated with when, why, and in what way… within an interactive work and social context.

Let your humanoid robots learn from the most accurate, real-world, 3D/4D model available… the human being in motion.

Wearable Technology. Wearing.

Now that smartphones are widely adopted, wearable devices, head-mounted display, and smart fabrics are the next big innovation, which will essentially place a persistent technology foundation on the human body that tracks and communicates passive and active activity information in real time. Whether the purpose is to stay fit and active, lose weight, be more organized, track your overall mental and physical health, or be entertained and communicate, the wearable will need to be strategically located on the body and designed to be worn comfortably and ergonomically empathetic with our daily activities. To achieve these objectives, 3dMD customer teams are using 3dMD Real-World data to ergonomically design the device to actual subject body types and calibrate wear with the subject’s natural movement and actions. The progressive sequence of 3dMD images conveniently captures a subject’s movement to evaluate overall dynamic flow while wearing the device, as well as a single frame pose that may indicate a design refinement would improve the performance, comfort, and experience. This connected, intelligent, and immersive application is a natural evolution for 3dMD based on the early work by 3dMD customer design teams in the area of advanced prosthetics, orthotics, and performance sportswear. Recently researchers have been evaluating the integration environment of Personal Protection Equipment (PPE) with smart headwear to provide safety combined with improved communication and information resources through XR projections. Wearable Technology includes… Smartwatches, Fitness Trackers, Sports watches, Head-Mounted Displays (VR headsets and AR smart glasses), Hearables, Consumer Health, Smart Clothing, etc.

Human Factors | Anthropometric | Size and Fit. Wearing.

The traditional mass production of garments has been based on the measurements of one average customer represented as a single fit (or size) model and the subsequent creation of other sizes based on proportional increases. This approach ignores the wide variety of body shape types within each size category thus creating garments that do not fit many consumers and causing supply chain inefficiencies for manufacturers and retailers. To address these discrepancies, 3dMD customer teams are actively capturing Real-World body type variations within a size category, as well as population samples with meaningful ethnic and age diversity, to modify their legacy sizing systems. Additionally, with the current trend for traditional brick-and-mortar retailers, 3dMD customer teams are evaluating local Make-to-Order opportunities. 3dMD systems provide workflow efficiencies that optimize the Episode of Encounter per consumer ensuring a positive experience with no physical contact. Whether it is producing well-functioning safety gear, developing a better fitting shoe or glove, improving performance sportswear and uniforms, or designing seats that improve comfort, 3dMD captures the individual consumer’s dynamics and Real-World anatomical 3D shape and size data precisely and economically. In 2016 a 3dMD customer team integrated 3dMD into the first pilot on-demand apparel production pipeline where the consumer is imaged with 3dMD upon arrival and the garment is subsequently manufactured prior to departure from the facility.

XR | AR | VR | MR. Training.

XR interactivity and immersion can truly enhance communications to create a memorable user experience whether it is a shopper trying on a watch on her actual wrist; a soldier being trained how to respond to a dangerous situation; or a worker connecting to a conference room surrounded by other virtual colleagues. To ensure that the user feels transported to another situation or sees products within his/her own room, visual design of the virtual environment and user experience is key, which brings us to XR’s biggest challenge… creating convincing virtual environments with humans. For some virtual environments, this will require the creation of life-like avatars that can participate with a user or a group of users for minutes or even hours. Sustainable engagements with others in the virtual environment will require avatars to effectively communicate and experience a range of emotions, which constantly sends signals from the face and body, in reaction to real-time stimuli and situations. To address this, 3dMD customer teams are building diverse Real-World human performance/emotion databases to use as avatar training sets. On a separate but related note, other 3dMD customer teams are building Real-World repositories of hand gesture information as this is proving to be an efficient way to control the UI. They are tracking hand motion along with finger articulation so users can reach out and interact with objects in the virtual environment.

Genetics | Biomarkers. Imaging.

Preventative medicine takes a proactive approach to patient care by establishing an individual’s health and well-being prior to the need for acute interventional treatment. To support this important endeavor in today’s world, 3dMD customer teams are working to identify biomarkers that can help flag possible disease risk, facilitate early detection, and assist diagnostic classification to guide more effective intervention and subsequent treatment. Supporting a personalized, preventive, and predictive approach, external Biomarkers related to the surface of the patient’s body are continuously being identified and studied. Over the years 3dMD customer teams in genetics and disease control research have studied and published about morphological anatomical structures that they believe may serve as a visual indication of a possible underlying condition. 3dMD images enable these research teams to define, identify, and interpret their own biomarker sets with the possibly of observing anatomical factors that through applied research could lead to aggregate improvements in early detection for patients and their families.