Image Processor Reducing Frame Delays in Digital Cameras
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image processing methods in digital cameras experience frame delays and high power consumption due to redundant processes, particularly when converting large-size RAW images into YCrCb data and reducing image sizes for display and recording, which can miss real-time display opportunities and increase power usage.
Innovation Solution
An image processor that directly generates multiple reduced-size RAW images from a large-size original image within a frame period, using horizontal and vertical reduction processors to produce images of arbitrary sizes for display, recording, and face detection, thereby reducing memory access and redundant processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If large-size RAW images are converted into YCrCb data and reduced to display size through conventional processing, then color image quality is maintained, but frame delay increases and power consumption rises
Solution Approach 1:
The patent applies preliminary action by generating multiple reduced-size RAW images directly from the original large-size RAW image before the conventional YCrCb conversion and display processing. This preliminary reduction step creates smaller image versions that can be processed more quickly, reducing frame delay while maintaining the option to generate full-size images when needed for quality requirements.
Solution Approach 2:
The patent segments the image processing workflow by creating separate reduced-size RAW images for different purposes (display, recording, face detection) from the original large-size RAW image. This segmentation allows each processed image to be optimized for its specific use case, reducing overall processing time and power consumption while maintaining quality where needed.
2Adaptability or versatility
If multiple image size reductions are performed sequentially after YCrCb conversion, then display and recording requirements are met, but power consumption increases due to redundant processes
Solution Approach 1:
The patent performs image size reduction as a preliminary action on the RAW image data before YCrCb conversion. By reducing the image size early when the data is still in its original format, the system avoids subsequent redundant reduction operations that would consume additional power. The reduced RAW images are then directly used for display and recording.
Solution Approach 2:
The patent makes the reduced-size RAW images universal by using them for multiple purposes: display processing, recording, and face detection. This multi-functionality eliminates the need for separate processing chains for each purpose, reducing redundant operations and overall power consumption while maintaining adaptability to different requirements.
3Manufacturing precision
If conventional image processing pipeline is used, then color accuracy is maintained, but processing speed decreases missing real-time display opportunities
Solution Approach 1:
The patent applies preliminary action by reducing the image size to display requirements before the YCrCb conversion and subsequent processing steps. This preliminary reduction creates smaller data sets that can be processed faster, increasing productivity and enabling real-time display opportunities while maintaining color accuracy through proper RAW image handling.
Solution Approach 2:
The patent introduces dynamics by allowing flexible selection of image sizes based on real-time requirements. The system can dynamically choose to process reduced-size images for real-time display or full-size images for quality-critical applications, adjusting processing speed and quality levels according to current operational needs.
Data Source
AI summary
A digital camera includes: a horizontal reduction (resizing) processor 301 for reducing a RAW image from single-sensor color imaging device to an image corresponding to a video recording size in an input line direction; a memory section 303 storing horizontally resized image data; a plurality of vertical reduction (resizing) processors 304, 306, and 308 reducing (resizing), in a vertical direction orthogonal to the input line direction, a plurality of pieces of reduced line data read out from the memory section 303; and horizontal reduction (resizing) processors 310 and 312 reducing a plurality of images reduced (resized) in the horizontal and vertical directions back into images of a display size and a face detection size in the input line direction.