Report: Computer Vision Teams Around The World Say Projects Are Delayed By Insufficient Data


Hear from CIOs, CTOs, and other executives and senior executives on data and AI strategies at the Future of Work Summit on January 12, 2022. Learn more


According to new research from Data generator, 99% of computer vision (CV) teams saw a machine learning (ML) project canceled due to insufficient training data. Delays, meanwhile, seem truly pervasive, with 100% of teams reporting significant project delays due to insufficient training data. Research also indicates that these training data challenges come in many forms and affect CV teams to an almost equal extent. Top issues faced by CV teams include poor annotation (48%), inadequate domain coverage (47%), and sheer rarity (44%).

The paucity of robust and domain-specific training data is only compounded by the fact that the field of computer vision lacks many well-defined standards or best practices. When asked how training data is typically collected in their organizations, interviewees revealed that a patchwork of sources and methodologies are used both in the field and within individual organizations. Whether synthetic or real, collected internally or from public datasets, organizations seem to use all the data they can to train their computer vision models.

However, computer vision teams have already identified and started adopting synthetic data as a solution. Ninety-six percent of CV teams said they have already embraced the use of synthetic data to help train their AI / ML models. Nevertheless, the quality, the source and the proportion of synthetic data used remain very variable from one domain to another, and only 6% of the teams currently use exclusively synthetic data.

This wave of synthetic data adoption is consistent with a number of recent industry reports predicting that 2022 will be a breakthrough year for synthetic data. This growing consensus certainly bodes well for the many highly anticipated applications of computer vision. In fact, it is possible that these technologies are much closer to their realization than they appear. Who knows? Maybe we are just a few good datasets from a driverless world.

The report is based on the findings of an online survey of 300 computer vision professionals representing 300 unique companies.

Read it full report by Datagen.

VentureBeat

VentureBeat’s mission is to be a digital public place for technical decision-makers to learn about transformative technology and conduct transactions. Our site provides essential information on data technologies and strategies to guide you in managing your organizations. We invite you to become a member of our community, to access:

  • up-to-date information on the topics that interest you
  • our newsletters
  • Closed thought leader content and discounted access to our popular events, such as Transform 2021: Learn more
  • networking features, and more

Become a member


Source link

Previous Buying a used GPU? Here's what to watch out for
Next How to Read Kindle Books on Computer Easily