
Enabling efficient Implementation
of neural networks in
smart cameras
Hosted by Francisco Socal
Summary
Artificial intelligence (AI), and neural networks (NN) in particular, is emerging as a key disruptive technology with a wide range of applications in different markets and segments. It enables a new class of smart cameras with advanced video analytics, abnormal behaviour detection, object and people recognition, to name a few applications. Implementing NNs in a camera SoC is particularly challenging due to the increased power consumption, memory bandwidth and silicon area required. This webinar provides a market overview of NNs in the smart camera space, including applications and trends.
It then focuses on the key SoC implementation challenges, technical requirements for a deployable and efficient solution and how these requirements can be addressed with innovative new hardware acceleration technologies.
Benefits of watching
- Gain an understanding of the smart cameras market trends and applications, including a look at the key players in the emerging ecosystem.
- Learn the key technical requirements and challenges for implementing a camera SoC for efficient AI and NN.
- Understand the trade-offs between CPU, GPU and neural network accelerator based implementations.
- Discover how new hardware acceleration technologies like the PowerVR 2NX Neural Network Accelerator can deliver the needed solution to meet performance, power, area and bandwidth requirements.
About the speaker
Francisco Socal
Product Manager for PowerVR Vision & AI, Imagination
Francisco Socal is a Product Manager for PowerVR Vision & AI at Imagination Technologies. Prior to joining Imagination in 2016, Francisco worked for London-based sports media and technology company Supponor Oy, first as a Quality Assurance Manager for two years, and then as Product Manager for five years. Prior to that, he was a Field Application Engineer and Customer Project Manager for On2 Technologies (now part of Google).