Video Slice Tiling With Implicit Entry Offsets for Lower Bit Overhead

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video coding techniques face inefficiencies in compression and processing overhead due to the use of slice headers and tile boundaries, which are unnecessary in modern network environments and complicate coding systems, wasting bits and reducing coding efficiency.

Innovation Solution

The solution involves partitioning video images into slices and tiles, ensuring each slice contains an integer number of tiles, computing the number of tiles based on the slice's corners, and omitting the end_of_slice_flag, thereby reducing bit overhead and simplifying the coding process.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If slice headers and tile boundaries are used in video coding, then processing and transmission can be organized into manageable units, but coding overhead increases and coding efficiency decreases

Engineering Contradiction:
Improvecoding efficiencyVSAvoidbit overhead
Core Design Contradiction:
ProductivityVSLoss of information

Solution Approach 1:

The patent removes the end_of_slice_flag from the slice header structure. By extracting and eliminating this redundant signaling element, the patent reduces bit overhead without compromising the ability to identify slice boundaries, as these can be determined through other means such as picture parameter sets or tile organization.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent modifies the slice header parameter structure by making the end_of_slice_flag implicit rather than explicit. This parameter change transforms a previously required signaling bit into a derived value, reducing the bit rate while maintaining the functional capability to process slice boundaries correctly.

Inventive Principle:
Principle #35Parameter changes

2Reliability

If slice headers are rewritten and processed, then slice boundaries can be identified, but processing overhead increases

Engineering Contradiction:
Improveslice boundary identificationVSAvoidprocessing overhead
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent establishes slice and tile boundary information in advance through picture parameter sets and tile organization structures. By pre-defining these boundaries before the actual slice header processing, the patent eliminates the need for real-time end_of_slice_flag checking, thereby reducing processing overhead while ensuring reliable boundary identification.

Inventive Principle:
Principle #10Preliminary action

3Adaptability or versatility

If entry point offsets are explicitly signaled for each tile, then random access to tiles is enabled, but bitstream size increases

Engineering Contradiction:
Improverandom access capabilityVSAvoidbitstream size
Core Design Contradiction:
Adaptability or versatilityVSLoss of information

Solution Approach 1:

The patent enables the decoder to self-determine tile entry point offsets by deriving them from the tile grid structure and picture parameter sets. Instead of explicitly signaling each offset, the system allows the decoder to calculate positions based on known tile dimensions and grid layout, maintaining random access capability while eliminating redundant signaling data.

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP4422175B1Slicing and tiling in video coding
Publication Date: 2025.11.05 HUAWEI TECH CO LTD
  • EP4422175B1 patent drawingFigure 1
  • EP4422175B1 patent drawingFigure 2
  • EP4422175B1 patent drawingFigure 3

AI summary

A video coding mechanism is disclosed. The mechanism includes receiving at a decoder, a bitstream including a video coding layer (VCL) network abstraction layer (NAL) unit containing a slice of image data divided into a plurality of tiles. A number of the tiles in the VCL NAL unit are determined. A number of entry point offsets for the tiles is also determined as one less than the number of the tiles in the VCL NAL unit. Each entry point offset indicates a starting location of a corresponding tile in the VCL NAL unit. The number of entry point offsets is not explicitly signaled in the bitstream. The entry point offsets for the tiles are obtained based on the number of entry point offsets. The tiles are decoded at the entry point offsets to generate a reconstructed image.