Neuromorphic Event-Based Vision Sensor Patents 2026
Neuromorphic Event-Based Vision Sensor Patents 2026
Event cameras detect per-pixel brightness changes asynchronously, delivering sub-millisecond latency and ≥120 dB dynamic range. Patent filings from 2011–2026 show rapid expansion from laboratory prototypes into autonomous vehicles, defense, and consumer electronics.
From Bio-Inspired Prototypes to Commercial Event Camera Systems
Neuromorphic event-based vision sensors, also called Dynamic Vision Sensors (DVS) or event cameras, fire asynchronous timestamped spikes per pixel when local luminance changes exceed a threshold. This eliminates temporal redundancy, yielding data rates orders of magnitude lower than conventional cameras while preserving high-speed scene dynamics at ≥120 dB dynamic range.
The technology spans five interrelated sub-domains within this dataset: pixel-level event detection hardware including CMOS-based DVS arrays and DAVIS hybrid designs; in-sensor and near-sensor processing using processing-in-pixel-in-memory (P2M) paradigms; sensor fusion architectures pairing event cameras with infrared focal-plane arrays, LiDAR, and IMU; algorithm and hardware co-design including spiking neural networks and FPGA implementations; and novel material platforms such as van der Waals heterostructures and memristive crossbars.
Three developmental phases are traceable across the retrieved dataset. The foundational phase (2011–2016) established algorithmic groundwork through early DVS, ATIS, and DAVIS sensor families. The development and deployment phase (2017–2022) saw Sony Advanced Visual Sensing AG’s direct memory control patent and Sensors Unlimited’s infrared-neuromorphic fusion filings signal the shift from academic prototypes to engineered products.
The maturation phase (2023–2026) shows expanding assignee diversity in retrieved records, with OmniVision, Qualcomm, GM Cruise Holdings, Raytheon, and multiple Chinese academic institutions all filing. In this dataset, 8 distinct patent-filing assignees are identified across 6 jurisdictions, with US filings dominating and CN filings growing in recency.
Patent Activity by Technology Cluster and Filing Period
Across the retrieved dataset, four core technology clusters account for the bulk of patent activity: event pixel array hardware, sensor fusion systems, neuromorphic computing integration, and event-based algorithm pipelines. Filing intensity accelerated markedly after 2019, with the 2023–2026 maturation phase showing the broadest assignee participation.
Patent Count by Technology Cluster — Retrieved Records
Sensor fusion and hybrid imaging systems represent the most patent-dense cluster in this dataset, followed by event pixel array hardware and neuromorphic computing integration, reflecting commercial focus on deployment-ready multi-modal architectures.
↗ Click bars to explorePatent Filings by Developmental Phase — Retrieved Records
Filing volume in this dataset rises steeply from the foundational phase (2011–2016) through the maturation phase (2023–2026), with the most recent period accounting for the greatest number of distinct assignees and application domains represented in retrieved records.
↗ Click bars to exploreKey Deployment Domains for Neuromorphic Event Camera Technology
Within the retrieved dataset, neuromorphic event-based vision sensors are being deployed or patented across at least four major application domains: autonomous vehicles, defense and surveillance, robotics and UAVs, and consumer electronics — each exploiting distinct sensor advantages including sub-millisecond latency, high dynamic range, or ultra-low power.
Autonomous Vehicles & Transportation
GM Cruise Holdings LLC filed a neuromorphic edge computing system for autonomous driving (US, 2024) featuring a collocated processing-memory platform for edge inference. Sony Advanced Visual Sensing AG’s Environmental Model Maintenance patents (US, 2022; CN, 2024) demonstrate event-sensor-guided fusion with depth sensors for dynamic object state prediction. Literature on real-time optical flow for vehicular perception with both low- and high-resolution event cameras further confirms deployment readiness.
Autonomous SystemsDefense, Surveillance & Space
Sensors Unlimited, Inc. holds a cluster of patents covering infrared-neuromorphic fusion for defense target detection and tracking across US, EP, and CN jurisdictions (2021–2025), including passive low-light imaging via compressive sensing. Raytheon Company’s NeRF-based object reconstruction patents (US, 2024–2025) address ISR 3D scene modeling from event streams. Neutron irradiation testing at Los Alamos Neutron Science Center demonstrated fast event sensor recovery under wide-spectrum neutron bombardment, validating use in spacecraft.
Defense & SpaceRobotics and UAV Systems
Event-driven vision on a neuromorphic chip for drone control demonstrates 3 orders-of-magnitude improvement in speed-vs-power trade-off over conventional image sensors, per retrieved literature. The REVIO range-and-event-based visual-inertial odometry approach reduces localization error by approximately 29% compared to event-only methods. Sony Group Corporation’s Event-Based Imaging System for Vehicle, Drone, and Robotic System (WO, 2024) trains SNNs using event data fused with motion reference sensors to predict object motion.
Robotics & UAVsConsumer Electronics & HCI
Qualcomm filed a low-power always-on face detection, tracking, and recognition system using event sensors (CN, 2026) targeting mobile device and AR/VR wearable scenarios, enabling the main processor to remain in sleep mode until relevant visual events occur. OmniVision Technologies’ dual always-on/major-pixel array architecture (US, 2025) similarly targets battery-powered consumer devices. Event-based near-eye gaze tracking beyond 10,000 Hz has been demonstrated in the literature, far exceeding the 300 Hz ceiling of conventional frame-based eye trackers.
Consumer ElectronicsLeading Patent Assignees in Neuromorphic Event-Based Vision — Dataset Snapshot
In this dataset, Sony Advanced Visual Sensing AG is the most prolific patent filer with at least 6 retrieved patents, followed by Sensors Unlimited, Inc. with at least 5 patents concentrated in infrared-neuromorphic fusion for defense applications. Filing activity in retrieved records spans 8 distinct assignees across 6 jurisdictions, with US filings most numerous and CN filings growing most rapidly in recency.
Top Assignees by Filing Count — Neuromorphic Event Cameras (Dataset Snapshot)
↗ Click bars to exploreSony Advanced Visual Sensing AG
In this dataset, Sony Advanced Visual Sensing AG is the most prolific filer with at least 6 retrieved patents spanning US (2019–2024), WO (2019), and CN (2024) jurisdictions. Core patent families cover the Event-Based Vision Sensor with Direct Memory Control (WO 2019, US 2021, US 2024), Environmental Model Maintenance using event sensors (US 2022, US 2025), and an always-on event-guided environmental state prediction architecture. Multiple US continuations filed through 2024 indicate active prosecution of foundational readout circuitry.
Switzerland / Japan (Sony Group)Sensors Unlimited, Inc.
Sensors Unlimited, Inc. holds at least 5 retrieved patents across US, EP, and CN jurisdictions filed between 2021 and 2025, all concentrated in infrared-neuromorphic fusion imaging and compressive sensing for defense applications. Key patents include Neuromorphic Vision with Frame-Rate Imaging for Target Detection and Tracking (US 2021, EP 2021, US 2023) and Neuromorphic Compressive Sensing in Low-Light Environment (US 2024, EP 2023, US 2025). Patents span active and pending status, indicating continued investment in dual-use defense and surveillance capabilities.
United StatesSix Emerging Directions in Neuromorphic Event Camera Innovation (2023–2026)
Based on filings from 2023–2026 in this dataset, emerging directions span NeRF-based 3D reconstruction from event streams, always-on consumer device integration, event-camera-based adaptive optics, and integrated sense-store-compute chip architectures.
NeRF and 3D Scene Reconstruction from Event Streams
Raytheon Company filed two US patents in 2024–2025 applying neural radiance field (NeRF) learning to event camera input streams for 3D and 2D object and scene reconstruction, leveraging the sensors’ motion sensitivity to improve NeRF quality in dynamic scenes. A related 2024 patent uses dual ML models on neuromorphic delta-image time series to generate high-resolution synthetic images from event camera inputs. This represents a convergence of event sensing and photorealistic 3D modeling with direct ISR and robotics applications.
Always-On Consumer Device Integration
OmniVision Technologies’ 2025 US patent introduces a dual-pixel-class array architecture with dedicated always-on row/column scanners enabling wake-on-motion functionality at ultra-low standby power. Qualcomm’s 2026 CN filing for low-power always-on face detection, tracking, and recognition using event sensors targets mobile and AR/VR wearable scenarios where the main processor remains in sleep mode until visual events occur. These filings confirm mass-market smartphone and IoT endpoint commercialization is a near-term target.
Event-Based Vision Sensors vs. Conventional Frame-Based Cameras
Click any row to explore further.
| Dimension | Event-Based Vision Sensor (DVS) | Conventional Frame-Based Camera |
|---|---|---|
| Temporal Resolution | Sub-millisecond latency; asynchronous per-pixel firing | Fixed frame-rate clock; limited by readout interval |
| Dynamic Range | ≥120 dB | Typically 60–70 dB for standard CMOS sensors |
| Motion Blur | Immune to motion blur due to asynchronous pixel operation | Susceptible to motion blur at high object velocities |
| Power Consumption | Ultra-low; no redundant full-frame readout | Higher; full frame read at every clock cycle regardless of scene change |
| Data Output | Sparse asynchronous event stream (position, polarity, timestamp) | Dense pixel array frame at fixed intervals |
| Spatial Resolution | Currently lower relative to high-end frame cameras; hybrid DAVIS designs partially address this | Higher spatial resolution available in commercial sensors |
| Gaze Tracking Rate | Beyond 10,000 Hz demonstrated in literature | Ceiling of approximately 300 Hz for conventional frame-based eye trackers |
| Radiation Tolerance | Fast recovery under wide-spectrum neutron bombardment demonstrated at Los Alamos Neutron Science Center | Standard CMOS more susceptible to radiation-induced degradation in space environments |
Frequently Asked Questions: Neuromorphic Event-Based Vision Sensor Patents
A neuromorphic event-based vision sensor (also called a Dynamic Vision Sensor, DVS, or event camera) is a bio-inspired imaging device where each pixel independently fires an asynchronous timestamped spike when local luminance changes exceed a threshold, rather than capturing full image frames at fixed intervals. This delivers sub-millisecond latency, high dynamic range (≥120 dB), and ultra-low power consumption by eliminating temporal redundancy.
In this dataset, Sony Advanced Visual Sensing AG is the most prolific filer with at least 6 retrieved patents (US 2019–2024, WO 2019, CN 2024), followed by Sensors Unlimited, Inc. with at least 5 patents across US, EP, and CN jurisdictions focused on infrared-neuromorphic fusion for defense, and Raytheon Company with at least 3 patents filed in 2024–2025 on NeRF-based reconstruction and virtual sensor generation.
Based on retrieved records, the primary application domains are: autonomous vehicles and intelligent transportation (optical flow, visual odometry, edge inference); defense, surveillance, and space (infrared fusion, low-light compressive sensing, ISR 3D reconstruction, space radiation tolerance); robotics and UAVs (visual-inertial odometry, SLAM, spiking neural network drone control); and consumer electronics and HCI (always-on face detection, AR/VR gaze tracking beyond 10,000 Hz).
Based on filings from 2023–2026 in this dataset, emerging directions include: NeRF-based 3D scene reconstruction from event streams (Raytheon, US 2024–2025); always-on consumer device integration (OmniVision US 2025, Qualcomm CN 2026); event-camera Shack-Hartmann wavefront sensing for adaptive optics (Soochow University CN 2025, Zhejiang University CN 2026); integrated sense-store-compute neuromorphic chips (Ningbo Institute of Materials Technology and Engineering CAS, CN 2025); reservoir computing for event processing (University of Michigan, WO 2025); and multispectral and non-visual-band event sensing (Sony Semiconductor Solutions, WO 2025).
Sony Advanced Visual Sensing AG’s direct-memory-control architecture, first filed in WO in 2019 with active US continuations through 2024, covers fundamental readout circuitry that couples an event pixel array with a memory controller directly writing per-event data (position, polarity, timestamp) to addressable memory. According to the strategic analysis in the content, IP strategists at competing sensor manufacturers must assess freedom-to-operate carefully before commercializing event-camera hardware, as this architecture is likely present in many commercial event sensor designs.
In retrieved records, multiple Chinese academic institutions have filed recent patents: Nanjing University holds an active US patent (2023) and a further US patent active in 2026 on novel neuromorphic vision system architectures. Soochow University (CN, 2025) and Zhejiang University (CN, 2026) filed event-camera-based wavefront sensing patents. The Ningbo Institute of Materials Technology and Engineering (Chinese Academy of Sciences) filed a CN 2025 patent on an integrated sense-store-compute neuromorphic chip. Chinese subsidiaries of global companies (Sensors Unlimited CN, Sony Advanced Visual Sensing CN, OmniVision CN, Qualcomm CN) are also filing in CN jurisdiction, consistent with China’s broader push in AI hardware and neuromorphic computing.
Data and insights on this page are based on a limited patent and literature dataset and are for reference only. Figures may not represent the complete technology landscape.