Nanoveu granted US patent to protect EMASS real-time object detection technology

Nanoveu (ASX: NVU) has been granted a key US patent to protect specialised solutions for real-time object detection and tracking developed by semiconductor business Embedded AI Systems (EMASS).

The exclusively licenced patent covers a novel non-maximum suppression (NMS) hardware block that measures only 0.06 square millimetres.

That is more than 90% smaller than typical solutions.

Lower energy costs

NMS has been designed to reduce the costs and complexity of EMASS’s proprietary chip technology.

It can handle up to 10,000 “bounding boxes” per second while operating under 1 megawatt to significantly extend device battery life and slash energy costs.

Bounding boxes are digital frames that isolate distinct objects within an image or video, enabling individual human faces to be distinguished in security footage or specific items to be highlighted in a retail setting.

Integral technology

NMS is integral to Vision AI, guiding applications such as smart security systems, autonomous navigation and augmented reality to prioritise clarity and identify only the most relevant objects while filtering out duplicates or overlaps.

EMASS’s hardware-accelerated approach runs Vision AI computations on dedicated hardware rather than general-purpose processors.

This reduces latency, cuts energy usage and shrinks the overall footprint, making it vital for edge devices such as drones, smart cameras and autonomous robots.

Reinforced commitment

EMASS founder Mohamed Sabry Aly said the US patent would reinforce the company’s commitment to efficient and edge-focused AI.

“By streamlining NMS at the hardware level, we can enable faster, smaller and more sustainable solutions,” he said.

“This milestone validates our approach and accelerates our ability to deliver transformative products to markets that demand high performance and minimal resource use.”

Vision AI adoption

Mr Aly said EMASS technology had the potential to advance the adoption of Vision AI across diverse markets, enabling smarter homes, safer cities and more sustainable operations.

“Unlike standard NMS methods that struggle under real-time conditions, EMASS’s patented design can track multiple objects simultaneously with lower energy draw,” he said.

“The result is a Vision AI system capable of performing complex tasks such as identifying several moving objects at once in real time and with minimal power usage, [which] has the potential to unlock new opportunities for augmented reality, smart homes and wearable AI products.”

EMASS will commence compatibility integration tests across various hardware and software ecosystems to confirm the technology’s transformative potential and publish a comprehensive report detailing energy consumption, latency and accuracy.