Renesas, Arcturus and Arm join us live to talk about the Smart City

The future of smart city analytics requires state-of-the-art software methods, specialized compute and efficient edge hardware to enable real-time tracking and edge AI analysis for applications such as pedestrian safety.

In this Tech Talk we will explore state-of-the-art tracking methods including motion prediction tracking, identity assignment, and reidentification techniques. The session will highlight techniques used to accelerate real-time performance and improve accuracy of current and future edge devices running Linux using Arm CPUs, Ethos-U NPUs and Mali GPU.

If you enjoyed this video, and don't want to miss any of our future Arm Tech Talks, please subscribe to our channel.

00:30 How can we get more involved in the discussion during today's Tech Talk?​
01:19 What upcoming topics will be covered in future Tech Talks?​
02:14 How is ARM addressing the advancements in smart city technology?​
07:08 What role do frameworks and open-source standards play in ARM’s ecosystem?​
09:23 What are the key factors driving smart city analytics at the edge?​
13:23 How does ARM's tracking algorithm achieve high performance in object tracking?​
26:45 Why was the ORB method chosen for feature matching?​
27:23 How does body part-based reidentification handle partial occlusions?​
27:59 What effect does segmenting visible body parts have on embedding scores?​
29:05 How does generative AI pose estimation aid in multicamera tracking?​
30:15 What are the key considerations for optimizing edge deployment pipelines?​
32:07 What challenges are faced in edge AI system implementation?

Stay connected with Arm: