Dynamic Image Capture Resolution for Low-Power Vision Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Computer vision techniques on mobile devices face challenges in balancing performance and power consumption, leading to inefficient use of resources and increased demand for complex image processing tasks.
Innovation Solution
Implementing a system that captures images at varying resolutions based on processing conditions, allowing initial low-power processing on wearable devices and offloading complex tasks to companion devices, thereby reducing resource usage and power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If complex image processing tasks are performed on mobile devices, then processing performance is improved, but power consumption increases
Solution Approach 1:
The image processing task is divided into two segments: a low-power assessment phase performed on the wearable device to determine feasibility, and a high-power processing phase performed on the companion device. This segmentation allows the system to avoid unnecessary high-power processing while maintaining the ability to perform complex tasks when needed.
Solution Approach 2:
A preliminary assessment is performed on the wearable device to evaluate whether an image can be captured with sufficient resolution for OCR processing. This preliminary action filters out images that would require excessive processing power, allowing the system to make informed decisions about where to perform the full processing task.
2Measurement precision
If image capture resolution is increased, then image processing accuracy is improved, but device resources and power consumption increase
Solution Approach 1:
The system dynamically adjusts the image capture resolution based on the processing conditions and device resources available. Instead of using a fixed high resolution, the system adapts the resolution to match the computational capabilities of the device, capturing images at the minimum necessary resolution for accurate processing.
Solution Approach 2:
The system changes the capture parameter (resolution) based on the processing condition assessment. When the assessment indicates that high resolution is not necessary, the system uses lower resolution to reduce resource consumption while maintaining sufficient accuracy for the task at hand.
3Speed
If computational tasks are performed onboard the device, then processing speed is improved, but device complexity and power consumption increase
Solution Approach 1:
The companion device serves as an intermediary that handles the computationally intensive processing tasks. The wearable device only performs the lightweight assessment and coordination functions, while the companion device performs the heavy lifting for image processing, effectively distributing computational complexity across two devices.
Data Source
AI summary
In one general aspect, a method can include capturing, using an image sensor, a first raw image at a first resolution, converting the first raw image to a digitally processed image using an image signal processor, and analyzing at least a portion of the digitally processed image based on a processing condition. The method can include determining that the first resolution does not satisfy the processing condition; and triggering capture of a second raw image at the image sensor at a second resolution greater than the first resolution.