Contextual image stabilization
Inventors
Wall, III, Conrad • Xu, Minnan • Rizzo, III, Joseph F.
Assignees
US Department of Veterans Affairs • Massachusetts Eye and Ear
Publication Number
US-9389431-B2
Publication Date
2016-07-12
Expiration Date
2032-11-05
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Abstract
Image stabilization systems and methods include a detector configured to detect images, an actuator coupled to the detector, a sensor coupled to the detector and configured to detect motion of the detector, and an electronic processor in communication with the sensor and the actuator, where the electronic processor is configured to, for example: (a) receive information about motion of the detector from the sensor; (b) determine components of the motion of the detector, and associate a class with each of the determined components; (c) identify components to be compensated from among the determined components based on the associated classes; and (d) generate a control signal that causes the actuator to adjust a position of at least a portion of the detector to compensate for the identified components.
Core Innovation
The invention features image stabilization systems and methods that detect and compensate for motion of a detector while acquiring images. The system includes a detector configured to detect images, a sensor coupled to the detector to detect its motion, an actuator coupled to the detector for positional adjustments, and an electronic processor in communication with both the sensor and actuator. The processor receives motion information, determines components of this motion while associating each component with a defined class, identifies which components require compensation, and generates control signals causing the actuator to adjust the detector position accordingly.
The problem being addressed arises particularly in prosthetic devices worn by patients, where involuntary motion leads to visual artifacts such as shakiness and blurring, which impair the usefulness of acquired visual information. Conventional approaches use software post-processing algorithms that demand increasing hardware capabilities, which grow in size and power consumption, thus being impractical for portable devices. The invention aims to provide real-time or near-real-time compensation by directly stabilizing image acquisition through physical adjustment of the detector, reducing reliance on heavy post-processing.
The methods incorporate context-based analysis to classify components of the motion as voluntary or involuntary, compensating only for involuntary motions (such as involuntary head movements during locomotion) while leaving voluntary motions (such as deliberate gaze shifts) uncorrected to respect wearer intent. This differentiation mimics natural visual stabilization mechanisms seen in humans, allowing the system to provide stabilized visual information appropriate to the circumstance.
Claims Coverage
The patent includes two independent claims covering an image stabilization system and a method for image stabilization. Each independent claim centers on the detection, classification, compensation, and control of detector motion.
Image stabilization system with context-based compensation
A system comprising a detector configured to detect images, a sensor coupled to the detector to detect motion, an actuator coupled to the detector, and an electronic processor. The processor receives motion information, decomposes the detector's motion into components, associates each component with a class, identifies which components are to be compensated based on their classes, and generates control signals to actuate positional adjustments of at least a portion of the detector to compensate for those identified components.
Method for image stabilization using motion classification and compensation
A method that includes obtaining image information via a detector, detecting motion of the detector during acquisition, determining motion components and associating each with a class, identifying components to be compensated based on these classes, and adjusting the position of at least a portion of the detector to compensate for the identified components while acquiring image information.
The independent claims cover an image stabilization system and method that actively detect and classify components of detector motion, selectively compensating for involuntary motion to improve image stability during acquisition. The features emphasize sensing, classification into motion classes, selective compensation via positioning actuators, and real-time adaptive control.
Stated Advantages
Reduces computing hardware requirements and power consumption compared to software post-processing, enabling implementation in portable applications.
Provides real-time or near-real-time motion compensation by circumventing feedback bottlenecks present in software-only image stabilization methods.
Selectively compensates for involuntary motions while preserving voluntary movements, delivering visual information that accurately reflects wearer intent.
Mimics natural human visual stabilization mechanisms, enhancing the quality of visual information for users, particularly those with reduced visual acuity.
Documented Applications
Visual prosthetic devices for persons with reduced visual acuity, including systems embedded in eyeglass frames or headwear.
Robotic vision systems, manufacturing and assembly line inspection systems, and automated vehicle guidance systems vulnerable to detector motion artifacts.
Head-mounted devices such as hats, caps, or helmets worn by soldiers, police officers, or athletes for enhanced stabilized visual information.
Sensory substitution and augmentation devices that process various types of measured information, not limited to visual signals, for assisting individuals with sensory imbalances or augmenting human perception.
Interested in licensing this patent?