SPAD Photon Detection for Low-Photon Object Classification

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing visual object detection and classification systems face challenges in low lighting conditions, slow computation speeds, high power consumption, and privacy concerns, particularly in scenarios requiring rapid and efficient object recognition.

Innovation Solution

A computing system utilizing a photon detection system with SPAD arrays and a low-photon-count classification model that stabilizes classification after receiving a low number of photons, employing a classification evidence vector and incremental updating to provide rapid and accurate object recognition.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Illumination intensity

If conventional visual detection systems are used in low lighting conditions, then more photons are required for detection, but this increases exposure time and power consumption

Engineering Contradiction:
Improvelighting conditionsVSAvoidexposure time
Core Design Contradiction:
Illumination intensityVSLoss of time

Solution Approach 1:

The patent changes the detection parameter from conventional intensity-based detection to time-of-flight measurement at the photon level. By measuring the arrival time of individual photons rather than accumulating photon intensity, the system achieves accurate depth and object detection in low-light conditions without increasing exposure time or power consumption

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent replaces conventional image sensors with SPAD (Single-Photon Avalanche Diode) arrays that detect individual photons and their arrival times. This substitution enables the system to function effectively in low-photon environments by counting individual photon events rather than requiring accumulated light intensity

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If conventional classification models are used, then accurate classification requires processing large amounts of visual data, but this increases computation time and power consumption

Engineering Contradiction:
Improveclassification accuracyVSAvoidprocessing speed
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent extracts only the essential features from photon data - specifically the time-of-flight measurements and spatial positions of individual photons. By working with this精简ed feature set rather than complete images or large volumes of visual data, the classification model achieves accurate object recognition with significantly reduced computation time and power consumption

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent segments the classification task into two stages: first, the SPAD array captures time-of-flight data for individual photons; second, a dedicated low-photon-count classification model processes this structured temporal and spatial data. This segmentation allows each component to be optimized for its specific function, improving overall processing efficiency

Inventive Principle:
Principle #1Segmentation

3Reliability

If conventional detection systems operate continuously to ensure reliability, then detection accuracy is maintained, but power consumption increases

Engineering Contradiction:
Improvedetection accuracyVSAvoidpower consumption
Core Design Contradiction:
ReliabilityVSUse of energy by moving object

Solution Approach 1:

The patent employs event-driven periodic action where the SPAD array and classification model are activated only when photon events occur or when detection is required. The system processes photons as they are detected rather than continuously acquiring and processing data, maintaining detection reliability while dramatically reducing average power consumption during operation

Inventive Principle:
Principle #19Periodic action

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

Enables accurate and efficient object classification with reduced photon count, faster processing times, and lower power consumption, suitable for low-light and time-critical applications while preserving privacy.

Implementation Method 1

a photon detection system including one or more cells each including one or more photon detectors that output a photon signature (e.g., an electrical signature) in response to photons being incident on the photon detectors

Methodology Applied
Scientific EffectPhotoelectric Effect: Photoelectric Effect

Data Source

PatentEP4147156B1System for low-photon-count visual object detection and classification
Publication Date: 2025.09.03 GOOGLE LLC
  • EP4147156B1 patent drawingFigure 1
  • EP4147156B1 patent drawingFigure 2
  • EP4147156B1 patent drawingFigure 3

AI summary

A computing system can be configured for low-photon-count visual object classification. The computing system can include a photon detection system including one or more cells. Each of the one or more cells can include one or more photon detectors. Each of the one or more photon detectors can be configured to output photon signatures in response to a photon being incident on the one or more photon detectors. The computing system can include one or more processors and one or more memory devices storing computer-readable data. The data can include a low-photon-count classification model and one or more instructions that, when implemented, cause the one or more processors to perform operations for low-photon-count visual object recognition. The operations can include obtaining a photon signature from a photon detection system. The operations can include providing the photon signature to a low-photon-count classification model. The operations can include determining, by the low-photon-count classification model, a classification of a visual object disposed in view of the photon detection system based at least in part on the photon signature. The operations can include providing the classification as output of the low-photon-count classification model.