At a glance
| Booth | T9404 |
| Country | US |
| Website | oculi.ai |
Company profile
Oculi develops AI vision around the OCULI Sensing and Processing Unit (SPU), which integrates AI Vision at the pixel level, runs AI directly on the chip, and transmits only the information needed to understand the scene. On-chip processing eliminates the data pipeline that drains battery, while decisions occur at the point of capture with no Cloud round-trip and live detection under 5ms for autonomous systems. Personally identifiable information is filtered before data leaves the device, with zero cloud dependency and all processing happening on-chip. Software Defined Vision Sensors™ let the same chip support applications ranging from wearables to defense and smart cities, and the SPU is fully programmable so the same hardware can serve deployments across industries. Its application examples include privacy-first urban monitoring, traffic intelligence and infrastructure analytics; robotics, autonomous vehicles and industrial automation; and ultra-low-power AI vision for consumer electronics, wearables, AR/VR and IoT edge devices. For unmanned and autonomous defense platforms, Oculi describes ultra low SWaP vision sensors and IntelliPixel® technology delivering real-time visual intelligence with complete edge autonomy and no network required, demonstrated at 30,000 FPS with less than 0.3W power consumption. OCULI holds exclusive worldwide rights to core technology developed with over $40M in research investment at Johns Hopkins University, plus a growing portfolio of in-house patents. Its leadership combines the original research behind the technology with commercial exits and decades of image sensor industry experience. Oculi is developing AI Vision semiconductors, software, and systems for Edge AI devices, Internet of Things (IoT), building monitoring and automation, transportation, robotics, and automotive markets. Its vision is to redefine machine vision by embedding intelligence at the pixel for real-time, energy-efficient, privacy-first AI Vision at the edge, and its mission is to enable devices to see with human-level efficiency for intelligent autonomous systems at global scale. Its technologies are designed with privacy and data protection in mind for adherence to GDPR, CCPA, and other relevant frameworks, and Oculi states a commitment to international laws and standards.
Exhibits
Most machine-vision alternatives use CMOS image sensors (CIS) like those in phone cameras, which are designed for high-resolution pictures rather than AI vision; Oculi instead captures only the information needed to understand what is in the scene. The OCULI SPU runs AI directly on the chip, integrates AI Vision at the pixel level, and transmits only the information needed to understand the scene. On-chip processing eliminates the data pipeline that drains battery and is purpose-built for always-on, battery-powered edge deployments. Decisions happen at the point of capture without a Cloud round-trip, with live detection under 5ms for autonomous systems, while personally identifiable information is filtered before data leaves the device and all processing happens on-chip with zero cloud dependency. Software Defined Vision Sensors™ let the same chip serve applications from wearables to defense and smart cities; the SPU is fully programmable so the same hardware can support deployments across industries. OCULI offers the solution as an OCULI SPU or complete Lens Module, and its Product Development Kit includes the tools needed to develop custom designs using Software Defined Vision Sensor(TM) capability. OCULI describes its solution as the world's first true AI vision solution, designed from the ground up to duplicate human vision for Physical AI with ultra-low power, low latency, high speed, low system-level cost, built-in privacy, and in-pixel computing at the chip level. Oculi leverages low-cost CMOS process technology and the CMOS Image Sensor ecosystem and implements the solution using a stacked Back Side Illumination (BSI) manufacturing approach. IntelliPixel® lets each pixel act as both a light-capture point and an intelligent processor, selectively capturing and processing data to reduce redundancy and transmit only pertinent information. Sensor-level processing reduces latency, bandwidth use, and power consumption for real-time and battery-operated applications, while on-sensor filtering supports privacy and programmability supports customization from industrial automation to consumer electronics. Applications include privacy-first urban monitoring, traffic intelligence, infrastructure analytics, robotics, autonomous vehicles, industrial automation, consumer electronics, wearables, AR/VR, and IoT edge devices. For unmanned and autonomous defense platforms, Oculi describes ultra-low SWaP vision sensors and IntelliPixel® technology delivering real-time visual intelligence with complete edge autonomy and no network required, demonstrated at 30,000 FPS with less than 0.3W power consumption. The Magenta Program simulates real sensor outputs for early-stage development without physical hardware and is intended for software teams building perception pipelines, object-detection models, or edge-AI inference engines; it also supports presence-detection data for privacy-conscious occupancy analytics, smart-building automation, and people-flow measurement where zero raw image data is required. Proof of Concept engagements are customized to explore specific applications and performance metrics, with Oculi's engineering team working with customers to define success criteria, deploy hardware, and validate results.
Capabilities and products
- pixel-level AI vision SPU
- on-chip AI vision processing
- edge live detection under 5ms
- personally identifiable information filtering before data leaves
- programmable Software Defined Vision Sensors™
- OCULI SPU and Lens Module supply
- 30,000 FPS vision sensing at less than 0.3W