Temporal De-noising via Persistent Image Buffer for Stereo Search
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies struggle to effectively reduce noise in images captured by cameras, leading to temporal instability and issues in applications requiring stable image comparisons, such as stereo correspondence detection.
Innovation Solution
The implementation of a computational/statistical model of the sensor to estimate noise and perform temporal de-noising operations, which involves evaluating a noise model, blending data from a persistent image buffer with current frame data, and updating the persistent image buffer to generate reduced-noise image frames.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing image processing technologies are used, then image processing can be performed, but noise reduction is insufficient leading to temporal instability
Solution Approach 1:
The patent applies preliminary action by maintaining a persistent image buffer that stores historical image data before current processing occurs. This pre-stored data is then used during temporal de-noising operations to compare and stabilize current frames, effectively preparing reference material in advance to combat noise and temporal instability.
Solution Approach 2:
The patent implements feedback through iterative temporal de-noising operations where the persistent image buffer is continuously updated with processed results. The system compares current frame data with historical data, applies noise reduction algorithms, and feeds the de-noised results back into the persistent buffer for future comparisons, creating a closed-loop system that progressively improves temporal stability.
2Object-affected harmful factors
If temporal de-noising operations are performed, then noise is reduced, but computational cost increases
Solution Approach 1:
The patent applies local quality by performing temporal de-noising operations selectively on specific image data rather than uniformly processing entire frames. The persistent image buffer stores and processes only relevant portions of historical data, and de-noising computations are focused on areas where temporal comparison provides the most benefit, reducing overall computational burden while maintaining effective noise reduction.
Solution Approach 2:
The patent uses copying by maintaining a persistent image buffer that replicates and stores historical frame data. Instead of re-processing original raw data, the system works with copied representations stored in the buffer, allowing efficient temporal comparisons and de-noising operations without the computational overhead of accessing and re-processing原始数据 repeatedly.
3Productivity
If image data is captured by camera, then image acquisition is achieved, but temporal instability and noise are introduced
Solution Approach 1:
The patent implements continuity of useful action by maintaining a persistent image buffer that continuously accumulates and retains historical image data across multiple frames. This continuous storage and progressive processing of temporal information allows the system to leverage ongoing data accumulation for noise reduction, transforming the continuous stream of captured images into a stabilizing resource rather than a source of temporal variability.
Data Source
AI summary
Methods, systems, devices and computer software/program code products enable generating reduced-noise image frames based on image frames received from a digital camera pipeline; and enable efficient stereo image search between corresponding images generated by at least two cameras.