Flash Detection in Video Compression Using Luminance Gradient Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
State-of-the-art video compression engines face poor compression performance when encoding video sequences with short, bright flashes, as existing flash detection methods are complex, costly, or inaccurate, leading to misleading scene change detection and artefacts.
Innovation Solution
A pre-processing method using normalized difference index values for luminance and chrominance components, combined with a luminance mean gradient index, to detect flashes with low computational complexity and high accuracy, suitable for implementation in low-cost coder equipment like VLIW digital signal processors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing flash detection methods are used, then flash detection capability is provided, but computational complexity and cost increase
Solution Approach 1:
The flash detection algorithm segments the video processing into distinct phases: pre-processing phase where flash detection occurs using simplified metrics, and main encoding phase where normal compression operates. This separation allows complex detection to be performed only when needed using lightweight operations, reducing overall computational burden while maintaining detection reliability.
Solution Approach 2:
The patent employs simple, low-cost detection metrics (normalized difference index, luminance mean gradient) that can be computed with minimal processing power. These lightweight detection mechanisms replace complex expensive algorithms, enabling flash detection in cost-effective hardware while consuming negligible computational resources during normal operation.
2Reliability
If existing flash detection methods are used, then flash detection is attempted, but accuracy decreases leading to false detections
Solution Approach 1:
The algorithm applies different detection strategies to different video characteristics by analyzing local properties of luminance and chrominance components. The normalized difference index specifically targets luminance changes while the luminance mean gradient captures temporal variations, allowing the system to adapt its detection sensitivity to local video conditions and avoid false detections in complex scenes.
Solution Approach 2:
The patent dynamically adjusts detection parameters including threshold values for normalized difference index and luminance mean gradient based on video content characteristics. By adapting these parameters to match the specific properties of each video sequence, the system maintains high detection accuracy across diverse scenarios while minimizing false positives.
3Speed
If flash detection is performed during encoding, then real-time detection is achieved, but processing power consumption increases
Solution Approach 1:
The flash detection algorithm performs preliminary analysis of video frames using simple metrics before the main encoding process begins. By detecting flashes in advance during a pre-processing stage, the system can prepare appropriate encoding parameters beforehand, enabling real-time response without consuming excessive processing power during the actual encoding operation.
Solution Approach 2:
The patent applies partial detection action by using simplified detection metrics that process only the most critical information (luminance components and their temporal variations) rather than performing complete complex analysis on all video data. This partial approach achieves real-time detection performance while consuming minimal processing power, applying enough analysis to detect flashes without excessive computation.
4Reliability
If scene change detection algorithms are used for flash detection, then detection is attempted, but artefacts are produced due to similar statistical characteristics
Solution Approach 1:
The algorithm creates asymmetric detection criteria that specifically target flash characteristics while avoiding symmetry with scene change detection. By using the normalized difference index combined with luminance mean gradient, the system establishes unique detection signatures for flashes that differ from typical scene change patterns, allowing discrimination between the two phenomena and preventing artefact generation.
Solution Approach 2:
The patent introduces intermediary detection metrics (normalized difference index and luminance mean gradient) that serve as mediators between raw video data and final detection decisions. These intermediary measurements provide a transformed representation of video content that highlights flash characteristics while filtering out misleading similarities to scene changes, enabling accurate distinction between flashes and normal scene transitions.
Data Source
Figure 1
AI summary
There is provided a method of pre-processing video data to detect flashes, comprising calculating a normalized difference index value for a luminance [Y] component of the video data, calculating a normalized difference index value for Chrominance [U and V] components of the video data, calculating a normalized luminance [Y] mean gradient index for the video data, and providing an indication that a flash has been detected when all the following conditions are satisfied: the normalized difference index value for the Luminance [Y] component is greater than a first predetermined threshold (TY); the normalized difference index value for the Chrominance [U and V] components is less than a second predetermined threshold (TC); and a difference between the normalized difference index value for the Chrominance [U and V] components and the normalized luminance [Y] mean gradient index is less than or equal to a third pre-determined threshold (TF).