vAIsual Inc, the company behind the largest visual dataset collection in the world, today launched its first biometric video dataset series.
vAIsual, established in 2021, will be licensing its collection of videos filmed specially for training machine learning.
The videos are shot on 4K resolution, and includes 1000 different individual identities with full body biometrics and postures.
Each model in the newly launched collection has signed a biometric model release, providing consent for the video and images to be used for machine learning.
According to vAIsual CEO, Michael Osterrieder, the greatest challenge facing the AI industry is the lack of compliance with existing laws around privacy and copyright.
“Our team comes from the professional stock media industry. We understand how to prepare and sell content in a legally clean way. We’ve used those skills to create AI training datasets that provide the legal clearances needed for commercial AI applications,” says Osterrieder.
The video datasets are an important step in ensuring companies are building AI technology in an ethical and responsible way.
“Cleanwashing has become the norm amongst generative AI companies who are trying to overcome the challenges of training their AI models on scraped data. We are the only company who can provide the full dataset disclosure that demonstrates 100% legal compliance,” says Osterrieder.
Video datasets are the most recent development on the datasetshop.com, which has thus far focused on still images for AI training.
According to Osterrieder, “Adding video opens up a whole range of options for AI teams, in particular, to learn action recognition. The video itself captures a huge amount of frames that can be extracted into separate still frames. These images provide a great variety of expression and movement.
“The datasets can be used for generative and discriminatory purposes, from automating the creation of films, to predicting movements and assessing health,” says Osterrieder.
The biometric image datasets are specially prepared for engineers to add to their workflow for AI training and are available via the Dataset Shop.
About Dataset Shop
First launched in 2022 by the “clean data guys”, vAIsual Inc’s Dataset Shop is a marketplace for visual media designed specifically for AI training purposes.
The online store initially sold the largest biometrically released human dataset, consisting of over 600,000 high quality images, custom shot for AI training.
The Dataset Shop is rapidly growing the collection of datasets through partnerships with stock agencies seeking to address the issue of widespread scraping of datasets, obtained without the consent of copyright owners.
About vAIsual (pro-nounced v-eye-sual)
vAIsual was first formed in 2020 by Michael Osterrieder and Nicolas Menijes, soon to be joined by industry veterans Mark Milstein and Istvan Novak. All founders are well connected to the IP licensing industry.
vAIsual covers the whole AI workflow, from dataset generation and delivery, to optimizing training sessions. They offer generated content to the commercial advertising industry as well as for the machine learning industry.
vAIsual exclusively relies on ethically sourced and legally clean datasets.