Device and Method for Encoding and Decoding a Plurality of Floating-Point Values

Simple SummaryContent extracted from patent full text and abstract with AI.

This patent describes a device and method for highly efficient lossless (and optionally lossy) compression and decompression of large sets of floating-point values. The approach is particularly suited for image data (such as HDR images) where each pixel's brightness is stored as a floating-point value. The innovation lies in predicting (estimating) the mantissa part of each floating-point value based on the values of neighboring data points (context), and then encoding only the difference (residual) between the actual mantissa and the predicted value. This context-based prediction significantly reduces the erratic behavior (entropy) of mantissa data, making it much more compressible and yielding better compression ratios.

Use CasesContent extracted from patent full text and abstract with AI.

• Lossless compression of scientific simulation or sensor data with floating-point values (e.g. satellite images, weather models, or physics simulations) to reduce storage and bandwidth requirements.
• Compressing high dynamic range (HDR) images and video streams for professional photography, cinematography, digital cinema, or film post-production.
• Efficient storage and transmission of floating-point based rendered scenes in computer graphics, mixed-reality rendering, and special effects.
• Archiving and transferring floating-point intensive datasets in medical imaging or industrial measurements.
• On-the-fly compression/decompression in real-time signal processing systems that utilize floating-point values.

BenefitsContent extracted from patent full text and abstract with AI.

• Achieves higher compression ratios than conventional floating-point compression methods by exploiting local correlations and context, sometimes enabling 2:1 compression ratios.
• Works in a lossless mode, allowing exact reconstruction of original data which is critical for scientific, medical, and professional imaging fields.
• Reduces storage costs and network bandwidth requirements for large floating-point datasets, enabling cost-effective data management and distribution.
• Is adaptable for both image and non-image floating-point data, including scientific, industrial, and graphics applications.
• Can be implemented in software or hardware, enhancing flexibility for integration into existing workflows and systems.
• Improves the compressibility of floating-point data without altering its precision or representation, preserving data integrity.

Technical Classifications (CPCs)

Inventors & Applicants

Patent Abstract

The invention relates to a device and a method for encoding and decoding a plurality of floating point values (F) by predicting the contained mantissa values on the basis of the context. The invention allows the very erratic behavior of the mantissas of typical floating point values to be attenuated while making encoding efficient. Exemplary embodiments describe a device for encoding a plurality of floating point values (F), comprising an extrapolation unit (110), a determination unit (120), and an encoding unit (130). The extrapolation unit (110) extrapolates floating point values (F) from a vicinity of a predetermined floating point value of the plurality of floating points values (F) in order to obtain an estimated value (S) for the predetermined floating point value (I). The determination unit (120) determines a residual value of the mantissa from a difference between the estimated value and the predetermined floating point value (I). The encoding unit (130) encodes the predetermined floating point value by encoding the exponent of the predetermined floating point value (F) and of the residual value of the mantissa.