Camera Firmware Customization for ML Enablement

Executive Summary

Our client is an American leader in wireless, enterprise mobility, and security camera solutions. The client’s product portfolio includes a cloud-managed family of cameras that have local storage to remove the need for NVRs and are based on the powerful processor to analyze video and provide valuable insights without the need to send those video files to the cloud. They wanted to enable advanced machine learning capabilities to detect objects and regions without adding any latency or deteriorating the video quality. They were looking for a technology partner who can help with a solution that can be deployed across different platforms with minimal resource consumption (CPU/GPU/DSP).

Leveraging strong experience in camera design and development, eInfochips proposed a Smart Codec solution that can detect motion and histogram from a frame on DSP to reduce CPU utilization and Smart Codec execution time. It reduced the on-edge storage and transmission bandwidth for Smart cameras without compromising the video quality.

Project Highlights

Camera Firmware Customization for ML Enablement
  • Firmware for Smart Codec
  • Enabled Advanced Analytics
  • Reduce the on-edge Storage and Transmission Bandwidth
  • Support Multiple Platforms Qualcomm SD624 and QCS610
  • High-Quality Video
To Download This Resource
Fill in the details below

Reference Designs

Our Work

Innovate

Transform.

Scale

Partnerships

Device Partnerships
Digital Partnerships
Quality Partnerships
Silicon Partnerships

Company

Products & IPs