Compressive-Sensing Image Capture With Optical Data Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional image sensors capture large amounts of data, which can be cumbersome and wasteful, as they often require mathematical reduction to manageability, consuming resources that could be used for processing.
Innovation Solution
A compressive-sensing image capture device with an array of sensors and passive or dynamic masks that block electromagnetic radiation, allowing only relevant data to be captured, reducing the need for processing and power consumption by performing filtering in the physical world.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If traditional image sensors capture all light reaching the sensor array, then complete image data is obtained, but large amounts of unnecessary data are produced that consume processing resources
Solution Approach 1:
The patent applies preliminary action by performing compression and filtering operations during the image capture process itself, rather than after capture. The compressive sensing mask is positioned in the optical path to pre-filter light before it reaches the sensor, encoding spatial information directly at the photon level. This preliminary encoding reduces the data volume at the source, eliminating the need to process and then discard large amounts of redundant pixel data later in the processing pipeline.
Solution Approach 2:
The patent extracts only the essential spatial frequency information from the optical field using a compressive sensing mask that selectively blocks and transmits light based on encoded patterns. Instead of capturing all spatial frequencies equally, the system extracts a compressed representation that contains sufficient information for image reconstruction while discarding redundant details at the capture stage, thereby improving processing efficiency without significant loss of reconstructable information.
2Quantity of substance
If mathematical compression is applied after image capture, then data reduction is achieved, but processing resources are wasted on handling large datasets
Solution Approach 1:
The system performs compression action preliminarily during light capture by using a compressive sensing mask to encode spatial information directly in the optical domain. This physical encoding reduces the data volume before it is converted to digital signals, meaning the sensor itself captures fewer photons and generates less data from the outset. Consequently, less energy is consumed during subsequent digital processing, ADC conversion, and data transmission, as the reduced data volume requires fewer computational operations throughout the entire pipeline.
Solution Approach 2:
The patent replaces mechanical/computational compression operations with an optical encoding mechanism. The compressive sensing mask uses physical light blocking and transmission patterns to encode spatial frequency information directly in the optical domain, substituting for later computational compression algorithms. This optical-domain compression occurs passively during light propagation, requiring no active processing power, thereby eliminating energy waste associated with post-capture mathematical compression of large datasets.
3Measurement precision
If full-resolution image data is captured and stored, then maximum image quality is maintained, but storage and processing requirements increase significantly
Solution Approach 1:
The system performs preliminary encoding of spatial information using a compressive sensing mask that captures spatial frequency content directly in a compressed form. By encoding the image in the spatial frequency domain during capture rather than in the spatial domain, the system maintains the essential structural information needed for high-quality reconstruction while storing far fewer data points. This preliminary frequency-domain encoding allows efficient storage of image essence without requiring full-resolution pixel data.
Solution Approach 2:
The patent transforms the image representation from the spatial dimension to the spatial frequency dimension through the compressive sensing mask. Instead of storing image data as a grid of pixel intensities in spatial coordinates, the system encodes the image as a compressed set of spatial frequency coefficients. This dimensional transformation allows the same image information to be represented more compactly, reducing storage requirements while preserving the information necessary for accurate image reconstruction through inverse transformation.
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
The device captures data that represents coefficients associated with different basis vectors or waveforms, reducing the amount of data processed and saving power and processing capacity, enabling efficient image analysis and reconstruction with simpler neural networks.
Implementation Method 1
Each sensor in the sensor array has an associated mask that blocks electromagnetic radiation from portions of the sensor
Implementation Method 2
The condensing lenses in some embodiments are metalenses that can compensate for small feature size of a pattern mask
Implementation Method 3
Other lens arrays may have refractive, diffractive, or metalens optical features that direct, filter, or change the phase of the incident wave
Implementation Method 4
A passive mask of some embodiments is a diffractive passive mask that comprises posts of different heights
Data Source
AI summary
Some embodiments provide a novel compressive-sensing image capture device and a method of using data captured by the compressive-sensing image capture device. The novel compressive-sensing image capture device includes an array of sensors for detecting electromagnetic radiation. Each sensor in the sensor array has an associated mask that blocks electromagnetic radiation from portions of the sensor. In some embodiments, an array of passive masks is used to block a particular set of areas of each sensor in the sensor array. In some embodiments, the image capture device also includes an array of lenses corresponding to the sensors of the sensor array such that each sensor receives light that passes through a different lens. Some embodiments of the invention provide a dynamic mask array. In some embodiments, a novel machine trained network is provided that processes image capture data captured by the compressive-sensing image capture device to predict solutions to problems.