Video Prediction Mode Selection With Bit and Memory Cost Balancing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video coding methods require massive video transmission from multiple cameras, leading to significant time delays in real-time analysis and decision-making, especially in applications like surveillance and intelligent transportation, and result in high power consumption due to complex coding modes and increased computing resources.
Innovation Solution
A video encoder and decoder system that employs memory-efficient prediction mode selection, optimizing cache efficiency by selecting coding modes that reduce memory reads and power consumption, while transmitting encoded video and extracted features separately for human and machine consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If conventional video coding methods are used to transmit video from multiple cameras, then video data can be transmitted, but power consumption increases and cache efficiency decreases
Solution Approach 1:
The patent changes the parameter of prediction mode selection by introducing memory cost as an additional criterion alongside bit cost. The encoder evaluates both bit cost (compression efficiency) and memory cost (cache efficiency) to select prediction modes that optimize both compression and memory access patterns, thereby reducing power consumption while maintaining real-time processing capability
Solution Approach 2:
The patent implements dynamic prediction mode selection that adapts to different processing conditions. The system dynamically adjusts the balance between bit cost and memory cost based on current operational requirements, allowing flexible optimization of power consumption and processing speed in real-time surveillance applications
2Loss of energy
If memory-efficient prediction mode selection is implemented, then cache efficiency improves and power consumption reduces, but encoding complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing cost values for different prediction modes. The encoder pre-evaluates both bit cost and memory cost for various prediction modes and stores these values for rapid comparison during encoding, avoiding complex real-time calculations and reducing overall encoding complexity
Solution Approach 2:
The patent introduces cost values as an intermediary metric that simplifies the decision-making process. Instead of directly optimizing multiple conflicting parameters, the system uses bit cost and memory cost as intermediate measures that guide prediction mode selection, making the complex optimization problem more manageable
3Loss of information
If prediction modes are selected based on bit cost only, then bitstream compression is optimized, but memory access efficiency deteriorates
Solution Approach 1:
The patent changes the selection criterion from considering only bit cost to considering both bit cost and memory cost. This dual-parameter approach ensures that prediction modes are selected to optimize both compression efficiency (bitstream quality) and memory access efficiency, preventing deterioration in either aspect
Solution Approach 2:
The patent creates a universal prediction mode selection mechanism that simultaneously optimizes multiple functions: bitstream compression and memory access efficiency. The dual-cost evaluation system serves both compression optimization and memory efficiency optimization, making the selection process multi-functional rather than single-purpose
Data Source
AI summary
A method of memory-efficient prediction mode selection includes receiving, by an encoder, a coded bitstream including a current frame, determining, by the encoder, costs of a first prediction mode and a second prediction mode, wherein determining further comprises determining, for the first prediction mode, a first bit cost and a first memory cost and determining, for the second prediction mode, a second bit cost and a second memory cost, selecting, by the encoder, a current prediction mode of the first prediction mode and the second prediction mode as a function of the first bit cost, first memory cost, second bit cost, and second memory cost, and encoding, by the encoder, the current frame using the current prediction mode. The prediction mode may be informed by at least one parameter received from the decoder.