Ocular system for deception detection
Inventors
Zakariaie, David • Bowden, Jared • Herrmann, Patricia • Weisberg, Seth • Sommerlot, Andrew R. • Anabtawi, Taumer • Brown, Joseph • Rowe, Alexander
Assignees
Publication Number
US-11928632-B2
Publication Date
2024-03-12
Expiration Date
2040-12-18
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Abstract
A method of deception detection based upon ocular information of a subject provides a video camera configured to record a close-up view of a subject's eye. A cognitive state model is configured to determine a high to a low cognitive load experienced by the subject. An emotional state model is configured to determine a high to a low state of arousal experienced by the subject. After asking a question, the ocular information is processed to identify changes in ocular signals of the subject. The cognitive state and emotional state models are evaluated based solely on the changes in ocular signals where a probability of the subject being either truthful or deceptive is estimated for a binary output.
Core Innovation
The invention is a method and system for deception detection based on ocular information from a subject. It employs a standoff device equipped with a video camera to record a close-up view of the subject's eye. A cognitive state model determines the subject's level of cognitive load, and an emotional state model assesses the subject's arousal, both based on ocular signals. After asking the subject a question, ocular information is processed to detect changes in these signals, allowing the system to estimate the probability that the response is truthful or deceptive, resulting in a binary output.
The problem addressed by the invention is the need for improved ocular systems capable of providing non-contact, objective, and immediate deception detection without reliance on traditional polygraph methods or the need for multiple instances or specific structuring of questions. Conventional approaches often require comparing results across several related questions or involve duplicate topics, reducing efficiency and flexibility.
The system uses sophisticated algorithms to analyze a range of ocular signals, including eye movement, gaze location, saccade characteristics, fixation metrics, pupil and iris features, blink dynamics, and more. These data are acquired and processed in real-time, enabling precise and rapid assessment of deception tied directly to individual responses. The entire process is contactless, and computational tasks can be performed locally or remotely, providing flexibility in deployment.
Claims Coverage
There are three independent claims in this patent, each focusing on a method of deception detection based on ocular information with specific inventive features.
Contactless ocular-based deception detection using a standoff device
A method utilizing a standoff device that does not physically contact the subject and incorporates at least one video camera to record close-up eye images of the subject. The method establishes a baseline of ocular information before questioning, records changes in ocular signals during and after a question, and processes this information to estimate a probability of truthfulness or deceptiveness using cognitive and emotional state models evaluated solely on ocular signal changes. The output is a binary result (truthful or deceptive) displayed to an administrator.
Weighted probability estimation over time for deception classification
A method in which the estimation of the subject being truthful or deceptive is performed by taking multiple probability estimates throughout the subject's response period. These estimates are weighted and combined to produce the final binary output, enhancing the robustness and temporal resolution of the deception assessment.
Saving binary output and corresponding ocular video data
A method that not only provides a binary output after deception detection but also explicitly saves each binary classification result along with the corresponding video recorded by the video camera. This approach creates a timestamped and retrievable record for each assessment, ensuring traceability and review of both the outcome and supporting ocular data for each session.
The independent claims cover a contactless, camera-based deception detection method leveraging models of cognitive and emotional state derived from ocular signals, with inventive features including baseline establishment, real-time and time-weighted probability estimation, and recording of results aligned with corresponding ocular video data.
Stated Advantages
Enables contactless and objective deception detection, improving flexibility and reducing intrusiveness compared to traditional systems.
Provides immediate, individual classification of responses without the need for duplicate or specially structured questions.
Facilitates high-fidelity ocular measurement and real-time data processing for accurate and rapid results.
Supports flexible deployment, including local or cloud-based computation and adaptable hardware setups.
Documented Applications
Deception detection during structured questions, interrogation, or passive observation of a human subject.
Assessment of a worker's operational risk, including fitness for duty, fatigue, intoxication, and psychological risk factors.
Optimization of learning and training by monitoring cognitive load, attention, engagement, and memory formation through ocular signals.
Inferring relationships between iris physiology and cognitive or emotional states using transillumination techniques.
Interested in licensing this patent?