Body Labs

Body Labs was a Manhattan-based software company specializing in computer vision and human-aware artificial intelligence. Its technology was designed to infer the shape, pose, measurements, and motion of a person from RGB photographs, video, or three-dimensional scans. Rather than treating a person as a collection of pixels, the company sought to create a computationally meaningful model of the human body that could be used by applications in areas such as digital apparel, animation, augmented reality, gaming, computer graphics, measurement, and interactive media. The company’s intellectual origins preceded its formal incorporation. Michael J. Black began investigating statistical models of the human body in 2002. During his work teaching computer vision at Brown University, a case involving surveillance footage of a robbery and murder at a 7-Eleven illustrated a practical use for body modeling. Law enforcement wanted to estimate physical characteristics of a suspect from video, and statistical modeling offered a way to assess information such as height. Black later became associated with William O'Farrell, Eric Rachlin, and Alex Weiss through Brown University and the Max Planck Institute for Intelligent Systems. Together, they formed Body Labs in 2013. Body Labs developed a software platform and application programming interfaces intended to make body understanding available to third-party developers and enterprise users. Its first major commercial offering, BodyKit, launched in 2015 as a collection of APIs and embeddable components for incorporating human-body representations into other software. The company subsequently introduced Body Labs Blue, which focused on taking physical measurements and estimating additional digital measurements for applications such as made-to-measure clothing. Body Labs Red addressed the conversion of three-dimensional scans into complete digital body models and included a partnership with 3dMD for processing scan data. The company also experimented with consumer-facing and media-oriented applications. Mosh, released for Apple iOS in 2017, used the company’s technology to estimate a subject’s three-dimensional pose and body shape and place visual effects on that subject. In the same year, Body Labs introduced SOMA, a human-aware AI system intended to infer three-dimensional body shape and motion from ordinary RGB photographs or video. The SOMA Shape API extended that capability to third-party applications by providing model and measurement prediction through a backend service. Body Labs raised $2.2 million in seed financing in November 2014, led by FirstMark Capital, with participation from New York Angels and existing investors. In November 2015, the company announced an $11 million Series A round led by Intel Capital, joined by FirstMark Capital, Max-Planck-Innovation GmbH, Osage University Partners, Catalus Capital, and the founders. These investments supported the development of its computer-vision platform and the expansion of its API and product portfolio. Amazon acquired Body Labs in October 2017. The acquisition brought Body Labs’ expertise in three-dimensional human modeling, pose estimation, and body-aware computer vision into Amazon. Public reference material does not establish a separate current operating status, independent website, or continuing standalone product lineup after the acquisition. Body Labs is therefore best understood as an acquired technology company and brand whose independent corporate trajectory became associated with Amazon after 2017.

Body Labs emerged from research into statistical representations of the human body. Michael J. Black began working on this subject in 2002, focusing on methods that could estimate physical form and movement from visual evidence. A practical example arose while Black was teaching computer vision at Brown University. Virginia State Police sought help analyzing surveillance footage connected with a robbery and murder at a 7-Eleven. Although video could not directly provide every physical fact about an individual, a statistical body model could help evaluate characteristics such as a person’s height and thereby support interpretation of the evidence. Black’s academic connections at Brown University and the Max Planck Institute for Intelligent Systems helped bring together the group that founded Body Labs: Black, William O'Farrell, Eric Rachlin, and Alex Weiss. The company was established in 2013 in New York and pursued the commercialization of research capable of reconstructing a three-dimensional representation of a person from ordinary RGB images and video. Its central proposition was that software could infer not only identity or object categories, but also the underlying geometry, pose, and movement of a human subject. Body Labs’ early business model emphasized developer tools and APIs. On March 3, 2015, it launched BodyKit, a set of APIs and embeddable components intended to let developers place human-body representations inside their own applications and creative tools. This approach gave the company potential reach across interactive entertainment, computer graphics, fashion technology, augmented reality, and other markets without requiring Body Labs to build every end-user application itself. The company expanded into apparel and measurement applications with Body Labs Blue, launched on July 20, 2016. Blue used the company’s body-understanding technology to take physical measurements and predict additional digital measurements. Such capabilities were relevant to custom clothing and other applications requiring a relationship between a person’s visual appearance and garment dimensions. On October 5, 2016, Body Labs launched Body Labs Red, an API for automatically turning three-dimensional scans into full digital body models. The company also announced a partnership with 3dMD, a provider associated with three-dimensional scanning, to process scan data. Body Labs continued to demonstrate its technology through consumer and media products. On February 15, 2017, it released Mosh through Apple’s App Store. The iOS application predicted a subject’s three-dimensional pose and shape and rendered effects over the subject, showing how the underlying computer vision could support an accessible mobile experience. On June 1, 2017, the company introduced SOMA, described as human-aware AI capable of predicting three-dimensional human shape and motion from RGB photographs or video. On July 21, it followed with SOMA Shape API, which exposed model and measurement prediction to third-party applications through the SOMA backend. Financing supported this development. Body Labs announced $2.2 million in seed funding on November 13, 2014, led by FirstMark Capital and joined by New York Angels and existing investors. On November 3, 2015, it announced an $11 million Series A led by Intel Capital, with participation from FirstMark Capital, Max-Planck-Innovation GmbH, Osage University Partners, Catalus Capital, and the founders. In October 2017, Amazon acquired Body Labs. The transaction was the decisive corporate turning point in the company’s history. It ended the period in which Body Labs operated as an independent venture-backed software company and placed its technology within Amazon. Publicly available reference material does not provide a complete account of the post-acquisition product roadmap, management structure, or whether the Body Labs name continued as a standalone commercial brand. Consequently, the company’s present independent status is recorded as unknown.

• 2017

  Mosh mobile application released

  [1]Body Labsen.wikipedia.org

  Body Labs released Mosh for iOS, using estimated three-dimensional pose and shape to apply visual effects to people in mobile content.

• 2017

  SOMA and SOMA Shape API introduced

  [1]Body Labsen.wikipedia.org

  Body Labs introduced SOMA for estimating three-dimensional human shape and motion from RGB imagery, followed by the SOMA Shape API for third-party model and measurement prediction.

• 2017

  Acquired by Amazon

  [1]Body Labsen.wikipedia.org

  Amazon acquired Body Labs in October, ending its period as an independent company.

• 2016

  Body Labs Blue launched

  [1]Body Labsen.wikipedia.org

  Body Labs launched an API and embeddable web interface for physical measurement capture and prediction of additional digital measurements.

• 2016

  Body Labs Red launched

  [1]Body Labsen.wikipedia.org

  The company introduced an API for converting three-dimensional scans into full digital body models and announced a partnership with 3dMD.

• 2015

  BodyKit launched

  [1]Body Labsen.wikipedia.org

  The company introduced BodyKit, a collection of APIs and embeddable components for integrating human-body representations into applications and tools.

• 2015

  Series A financing announced

  [1]Body Labsen.wikipedia.org

  Body Labs announced an $11 million Series A led by Intel Capital.

• 2014

  Seed financing announced

  [1]Body Labsen.wikipedia.org

  Body Labs announced $2.2 million in seed funding led by FirstMark Capital.

• 2013

  Body Labs founded

  [1]Body Labsen.wikipedia.org

  Michael J. Black, William O'Farrell, Eric Rachlin, and Alex Weiss founded Body Labs after academic and research connections involving Brown University and the Max Planck Institute for Intelligent Systems.

• 2002

  Early statistical human-body research

  [1]Body Labsen.wikipedia.org

  Michael J. Black began researching statistical models capable of representing the human body and estimating physical characteristics from visual information.

Body Labs positioned itself as a specialist provider of human-aware computer vision: software that could understand people as three-dimensional bodies with measurable shape, pose, and movement rather than merely recognize visual patterns in images.

• Developer platform and APIs · 2015

  BodyKit

  BodyKit was a collection of APIs and embeddable components for adding human-body representations to applications and creative tools. It reflected Body Labs’ developer-oriented strategy, allowing outside teams to use body-aware computer vision without building the underlying three-dimensional modeling system themselves.

• Body measurement software · 2016

  Body Labs Blue

  Body Labs Blue combined an API with an embeddable web interface for taking physical measurements and predicting further digital measurements. The product was particularly relevant to custom clothing and other use cases in which visual body information needed to be translated into apparel or product dimensions.

• 3D scan processing software · 2016

  Body Labs Red

  Body Labs Red was an API designed to process three-dimensional scans into complete digital body models. Its launch was accompanied by a partnership with 3dMD to process 3D scan data, extending Body Labs’ capabilities beyond ordinary photographs and video.

• Mobile augmented-reality application · 2017

  Mosh

  Mosh was an iOS application that predicted a subject’s three-dimensional pose and body shape and rendered visual effects on the subject. It served as a consumer-facing demonstration of Body Labs’ ability to interpret human form and movement from mobile imagery.

• Human-aware artificial intelligence · 2017

  SOMA

  SOMA was Body Labs’ human-aware AI software for predicting three-dimensional human shape and motion from RGB photographs or video. It represented the company’s broader platform vision: converting ordinary visual input into a structured, animated digital body model.

• Developer API · 2017

  SOMA Shape API

  SOMA Shape API exposed Body Labs’ model and measurement prediction capabilities to third-party applications through the SOMA backend. It was intended to make three-dimensional body reconstruction and measurement functionality easier for developers to integrate.

• 2017

  Launch of SOMA platformProduct launch

  Background

  Body Labs sought to consolidate its research and commercial capabilities around interpreting human shape and motion from ordinary visual input.

  Action taken

  The company launched SOMA and later released SOMA Shape API for three-dimensional model and measurement prediction.

  Aftermath

  SOMA became the clearest expression of Body Labs’ human-aware AI positioning shortly before the Amazon acquisition.

  Source

• 2017

  Acquisition by AmazonM&A

  Background

  Body Labs had developed a portfolio of computer-vision technologies for three-dimensional body modeling, pose estimation, and measurement prediction.

  Action taken

  Amazon acquired Body Labs in October 2017.

  Aftermath

  Body Labs ceased to be documented as an independent venture-backed company, and public reference material does not establish the subsequent standalone status of its products or brand.

  Source

• 2016

  Expansion into digital body measurementProduct launch

  Background

  Body Labs’ body-shape technology could support apparel and custom-fitting use cases requiring digital estimates of a person’s physical dimensions.

  Action taken

  The company launched Body Labs Blue, combining an API and embeddable web interface for physical and predicted digital measurements.

  Aftermath

  The launch broadened Body Labs’ commercial scope beyond general computer vision toward clothing and measurement applications.

  Source

• 2016

  Introduction of 3D scan processingProduct launch

  Background

  Professional three-dimensional scans provided richer body data than ordinary RGB images but required software capable of turning scans into usable models.

  Action taken

  Body Labs launched Body Labs Red and announced a partnership with 3dMD for processing three-dimensional scans into full body models.

  Aftermath

  The move extended the platform into professional scanning and digital-model workflows.

  Source

• 2015

  Launch of BodyKitProduct launch

  Background

  Body Labs needed a way to commercialize its human-body modeling research across multiple software categories and customer applications.

  Action taken

  The company released APIs and embeddable components that allowed developers to integrate human-body representations into their own products.

  Aftermath

  BodyKit established the company’s developer-platform orientation and was followed by specialized products for measurements, scan processing, and human motion.

  Source

1. 1Body Labsen.wikipedia.org