Anamorphic Lens Image Processing for Smart Glasses
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current computer vision applications face challenges in utilizing non-rectilinear image data from anamorphic lenses due to the need for re-training models, which is time-consuming and expensive, and these lenses introduce unique distortions that complicate processing, especially in applications like smart glasses where ergonomic considerations are crucial.
Innovation Solution
Employing anamorphic lenses with computer vision models that capture user interactions, utilizing non-square photosites and image signal processing to handle the optical squeeze and preserve the original light information, allowing for direct processing without conversion to rectilinear formats, and optimizing camera orientations to enhance field-of-view and comfort in smart glasses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If anamorphic lenses are used to capture images, then field-of-view and ergonomic comfort are improved, but image processing complexity increases due to optical distortions
Solution Approach 1:
The patent pre-calculates and stores distortion correction maps for various anamorphic lens configurations during system initialization. These pre-computed correction parameters are then directly applied during real-time image capture without requiring complex on-the-fly calculations, thus reducing processing complexity while maintaining the wide field-of-view benefits of anamorphic lenses
Solution Approach 2:
The patent introduces an intermediary processing layer that converts anamorphic lens output into a standardized rectilinear format compatible with existing computer vision models. This intermediary conversion layer acts as a mediator between the anamorphic optical system and standard processing pipelines, simplifying the overall system architecture by isolating the distortion handling to a dedicated conversion module
2Loss of information
If non-rectilinear image data from anamorphic lenses is used, then unique optical information is preserved, but computer vision model re-training is required which is time-consuming and expensive
Solution Approach 1:
The patent creates a digital copy or mapping relationship between anamorphic image coordinates and their corresponding rectilinear equivalents using pre-computed transformation matrices. This copying approach allows existing rectilinear computer vision models to process anamorphic images by simply applying coordinate transformations, eliminating the need for costly and time-consuming model re-training while preserving all optical information from the anamorphic lens
Solution Approach 2:
The patent transforms the problem from changing the computer vision model parameters to changing the image parameter representation. By applying parameter transformations to the image data itself (coordinate systems, distortion corrections) rather than re-training model parameters, the system maintains compatibility with existing models while utilizing anamorphic optical information
3Adaptability or versatility
If image conversion to rectilinear formats is performed, then compatibility with existing models is improved, but processing time and power consumption increase
Solution Approach 1:
The patent segments the image processing pipeline into distinct modular stages: anamorphic distortion correction, coordinate transformation, and region-of-interest extraction. Each segment is optimized independently with lightweight algorithms, allowing parallel processing and reducing overall processing time while maintaining compatibility with existing rectilinear-based computer vision models
4Loss of information
If full image processing is performed, then complete information is analyzed, but power consumption increases which is critical for smart glasses applications
Solution Approach 1:
The patent extracts and processes only the essential regions of interest from anamorphic images using pre-computed transformation maps that identify relevant areas. By extracting only necessary portions of the image for processing rather than analyzing the entire high-resolution anamorphic image, the system maintains information analysis completeness for relevant elements while dramatically reducing power consumption for smart glasses applications
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 efficient processing of anamorphic images without re-training computer vision models, reduces power consumption, and improves ergonomic hand positioning for comfortable user interactions in smart glasses, while maintaining the straightness of motion and lines, thus enhancing real-time applications like AR/XR and gesture detection.
Implementation Method 1
Anamorphic lenses have a cylindrical shape and optically squeeze light information along a major axis of the lens
Data Source
AI summary
Novel applications for anamorphic lenses in smart glasses. Anamorphic lenses preserve straightness of motion and lines (i.e., “linearity”). As a practical matter, existing computer vision models can be used with anamorphic images without re-training or intermediate conversion steps (unlike fisheye lenses). The contents of the present disclosure provide substantial improvements for applications that have different FOV requirements along different axis. Solutions for ergonomic hand placement (relative to gaze), non-square photosites, binning and eye-tracking are discussed throughout.