Veritone® Launches AI-Based Object Detection And Tracking Solution For Developers To Rapidly And Accurately Detect Objects In Visual Media

Without Requiring Data Science Expertise, Objectbox™ Provides Developers a Means to Easily Train and Deploy Highly Customized Cognitive Engines

COSTA MESA, Calif. – FEBRUARY 6, 2019 – Veritone Inc. (Nasdaq: VERI), the creator of the world’s first operating system for artificial intelligence, aiWARE™, today announced the launch of Objectbox, aiWARE’s newest cognitive engine in the Veritone Machine Box suite. Objectbox is state-of-the-art machine learning technology that developers can use to train their own custom object detection engines.  It is a unique tool in the market, accessible to those without data science expertise, combining advanced object detection capabilities with a user-friendly interface that allows users to get immediate feedback on detection efficacy. Objectbox has applications across all industries utilizing video and image content.  It can be deployed with aiWARE in the cloud or hybrid environments without the need to send data outside an organization’s firewall.

Object detection is a rapidly growing part of the computer vision industry, which is projected to grow to $48.6 billion by 2022. Objectbox can quickly detect selected objects with precision, saving users valuable time and resources. In the media and entertainment industry, for example, a broadcaster using Objectbox within aiWARE can inform its advertising clients when their logo appeared during a sporting event broadcast, how long it appeared on screen and how big the logo was, and can provide audience analytics — valuable information that helps the broadcaster demonstrate their client’s ad spend ROI.  Unlike many cognitive engines, Objectbox can run locally, avoiding the need to publish the user’s content to the cloud, which helps to address many privacy or legal concerns. This is of particular importance for government, legal and law enforcement users relying on the technology to aid efforts like eDiscovery, surveillance and suspect identification.

“At Veritone, we are democratizing AI and believe that it will play an increasingly larger role in helping to solve some of our world’s most complicated challenges.  Our unique AI operating system, aiWARE, together with the hundreds of powerful cognitive engines like Objectbox, and the growing list of applications on the platform, provides a configurable AI infrastructure that is steadily changing how people and organizations understand and extract valuable insights from their ever-growing volumes of data ,” said Chad Steelberg, Veritone Co-Founder and CEO. “Augmenting our AI platform with Objectbox offers our customers an incredibly easy-to-use solution for culling through vast amounts of video and imagery to find the objects they are seeking faster, accelerating and augmenting their business-critical workflows.”

How Objectbox Works

Objectbox allows developers to train detectors with their own content containing the desired objects to be detected, then deploy the resulting customized cognitive engine within aiWARE.  Accuracy can be extremely high because the developer can use samples from their own video and image data to train the detector via a simple user interface, which provides immediate feedback and allows for additional training to further improve detection accuracy, if desired.  After training against a few examples of objects in each scenario, Objectbox learns what that object looks like and can detect it in future images or video with great precision. For example, users can quickly and easily add a new brand or logo to their library, and train the engine to identify partial faces or body parts or highlight only certain sections of a video frame such as a scoreboard or news ticker.  

Once trained on a customer’s data, Objectbox is easily deployable. The developer downloads the training file with one-click in the user interface and stores it in a location of their choice with a link accessible to aiWARE.  When combined with aiWARE, Objectbox users will benefit from a production-ready environment with object detection output leverageable in an AI-enabled application or custom workflow. With the Veritone Ecosystem, users have the option to use Objectbox or a third-party engine partner for object detection based on their use case.  

Objectbox is the first new cognitive engine released by Machine Box for aiWARE since its acquisition by Veritone in September 2018. Two of Veritone’s newest industry applications, Veritone IDentify and Veritone Redact, are already leveraging Objectbox, enabling users to identify objects and faces in visual media rapidly and with ease.

About Objectbox

Objectbox is aiWARE’s newest cognitive engine for object detection and can be easily trained with developers’ specific data sets to yield greater accuracy.  Key benefits include:

  • Ease of use — Simple to train on customer data and easy to deploy regardless of skill level or knowledge.
  • Fast to market — Can be downloaded, trained and deployed in minutes.
  • Accuracy through trainability — User interface allows users to train Objectbox with use case-specific data sets and receive immediate feedback.
  • Portable deployment — Multiple deployment models available for cloud and hybrid environments.

To see a demo of the product, watch the video:

Objectbox will be made available as a part of the Veritone Cognitive Engine Ecosystem for existing and new aiWARE accounts to leverage.  Developers can trial Objectbox in minutes for free at machinebox.io/docs/objectbox. To get started, contact Veritone about deploying Objectbox within the aiWARE enterprise-ready production environment at unlock.veritone.com/machine-box-enterprise.


Safe Harbor Statement
This news release contains forward-looking statements, including without limitation statements regarding the expected growth of the computer vision industry and the expected benefits to customers of Veritone’s Objectbox technology. Without limiting the generality of the foregoing, words such as “may,” “will,” “expect,” “believe,” “anticipate,” “intend,” “could,” “estimate” or “continue” or the negative or other variations thereof or comparable terminology are intended to identify forward-looking statements. In addition, any statements that refer to expectations, projections or other characterizations of future events or circumstances are forward-looking statements. Assumptions relating to the foregoing involve judgments and risks with respect to various matters which are difficult or impossible to predict accurately and many of which are beyond the control of Veritone. Certain of such judgments and risks are discussed in Veritone’s SEC filings. Although Veritone believes that the assumptions underlying the forward-looking statements are reasonable, any of the assumptions could prove inaccurate and, therefore, there can be no assurance that the results contemplated in forward-looking statements will be realized. In light of the significant uncertainties inherent in the forward-looking information included herein, the inclusion of such information should not be regarded as a representation by Veritone or any other person that their objectives or plans will be achieved. Veritone undertakes no obligation to revise the forward-looking statements contained herein to reflect events or circumstances after the date hereof or to reflect the occurrence of unanticipated events.