Geometric Image Transforms for Axis-Aligned DCT Compression

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image compression methods result in lower compression ratios when compressing images with texture features that are not aligned with the image axis, leading to inefficient use of bits due to non-zero DCT coefficients.

Innovation Solution

Rotate or geometrically transform texture features in images to align them with the image axis before performing discrete cosine transform (DCT), generating mostly zero DCT coefficients for improved compression.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Area of moving object

If image compression is performed without aligning texture features, then the compression process is simple and fast, but the compression ratio is low due to non-zero DCT coefficients

Engineering Contradiction:
Improvecompression ratioVSAvoidcomplexity of alignment process
Core Design Contradiction:
Area of moving objectVSDevice complexity

Solution Approach 1:

The patent applies preliminary action by performing geometric transformation (rotation) on image blocks before conducting DCT compression. The system calculates the main direction within each block and rotates the block to align texture features with the horizontal/vertical directions before compression, thereby reducing non-zero DCT coefficients and improving compression ratio.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent changes the orientation parameter of image blocks through geometric transformation. By rotating blocks based on calculated main directions, the texture features are realigned to optimize the distribution of DCT coefficients, transforming the spatial orientation parameter to achieve better compression efficiency.

Inventive Principle:
Principle #35Parameter changes

2Loss of substance

If geometric transformation is applied to align texture features, then compression ratio improves, but processing time and computational complexity increase

Engineering Contradiction:
Improveredundant informationVSAvoidprocessing time
Core Design Contradiction:
Loss of substanceVSLoss of time

Solution Approach 1:

The patent segments the image into multiple blocks and processes each block independently. By dividing the image into smaller units, the geometric transformation and DCT operations can be performed on individual blocks, reducing the overall computational complexity and processing time compared to transforming the entire image at once.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies partial action by performing geometric transformation only on blocks where it is most beneficial. The system calculates the main direction for each block and applies rotation selectively, rather than uniformly transforming all blocks, thereby optimizing the balance between compression improvement and processing overhead.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3753250B1Geometric transforms for image compression
Publication Date: 2025.08.27 GOOGLE LLC
  • EP3753250B1 patent drawingFigure 1A~1B
  • EP3753250B1 patent drawingFigure 2A~2B
  • EP3753250B1 patent drawingFigure 3A

AI summary

A decoder may perform a method of decompressing images that include texture features that are not aligned with an axis of the image being compressed. In some example implementation, the method may include receiving a block of geometrically transformed pixel values and performing an inverse geometric transformation on the block of geometrically transformed pixel values to generate a first block of pixel values. The geometrically transformed pixel values represent texture features of an image that are non-parallel with a vertical axis or a horizontal axis of the image and the first block of pixel values being one of a plurality of blocks of the image. The example method may further include generating at least a portion of the image based on the first block of pixel values.