Case Studies

Computer Vision for Monitoring Breathing Rate in a Smart Baby Monitor

Highlights

Using the Chrysalis platform, CocoonCam was able to reduce costs to under $1 per user per month while still processing 15 PetaBytes of video and audio data per month, replacing the previous Kinesis/Fargate/Cloudwatch stack that resulted in sky-high monthly fees. In addition, Chrysalis enabled CV teams to build and test algorithms faster and ship new features in an accelerated timeline.

Overview

The CocoonCam was launched in 2018 and quickly became Amazon’s #1 smart baby monitor. This monitor used computer vision to monitor breathing rates, movement and crying in babies. As a result of high customer engagement, CocoonCam’s cloud computing costs spiralled and approached $70 per user per month. CocoonCam streamed 720p video at 10FPS in H.264 format. To continue scaling, the company needed to act quickly to reduce streaming and compute costs.

Conclusion

This case study proved Chrysalis’ value for customers who want to provide real-time insights to customers using advanced computing techniques like AI/ML. Cocoon was able to use an edge-cloud hybrid system to offer a product with low upfront cost while still while still maintaining the ability to offer new analytics services easily.