Building Virtual Reality, One Pixel at a Time

Challenge

This unicorn machine vision startup faced a large task: how to get confidential images tagged, segmented, and categorized quickly and discretely. They needed a solution that was not only accurate, but secure and seamless from end-to-end. While anonymous crowds provided speed, they did not guarantee security, nor were they flexible enough to fit the client’s workflow. In-house options, while secure, were expensive and slow.

Solution

iMerit development teams built a platform and workflow specifically for this client’s needs, and trained teams of workers to securely and quickly annotate client images exactly per their requirements.

As new projects – with new requirements – rolled in, iMerit workers were able to quickly adapt their previous learnings and tackle new challenges. We now work with images of facial features, real-world captures, and human gestures.

Results

The initial accuracy of iMerit teams was so impressive that the client immediately scaled the project to full production mode. In the initial project phases, iMerit teams segmented nearly 35,000 complex images in two weeks, and we have not stopped yet! Each day, hundreds of accurately tagged image files flow from iMerit into this client’s systems, ensuring that they build robust algorithms and build the next generation of machine vision.

Client

Unicorn Machine Vision Startup

Service

Image Segmentation, Annotation + Tagging, Dataset Creation

Vertical

Financial Services


RELATED CASES

Share this post with your networks:Share on FacebookTweet about this on TwitterShare on LinkedIn