Method and Devices for Processing Input Signals

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

This invention relates to a novel method and devices for efficiently processing input signals (such as images, audio, or video). The core idea is to decompose the input signal into a set of 'kernels'—basic analytical building blocks—with each kernel described by a set of parameters (like position, size, orientation, and intensity). These parameters are then quantized, coded, and used to reconstruct the signal or process it further. The approach works for multi-dimensional signals and allows for efficient coding, transformation, signal feature extraction, and high-quality signal reconstruction, including operations such as upscaling or noise reduction.

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

• Image and video compression (e.g., in multimedia streaming, broadcasting, or storage applications).
• Audio and speech compression and enhancement for telecommunication or storage.
• Super-resolution imaging (increasing the resolution of images or videos, for example, in medical imaging or satellite imagery).
• Noise reduction in audio, images, or videos.
• Feature-based signal retrieval or classification, such as content-based image retrieval (CBIR), automatic tagging, or similarity search systems.
• Efficient transmission and storage of multi-dimensional signals like lightfield images or videos.
• Adaptive rate conversion (scaling resolution or framerate) of multimedia content.

BenefitsContent extracted from patent full text and abstract with AI.

• More efficient signal compression, leading to reduced storage and transmission bandwidth requirements.
• Flexible application to any type of N-dimensional signal (including images, audio, video, or lightfield data).
• Ability to extract meaningful signal features directly from the compressed domain, facilitating tasks like classification and retrieval.
• High reconstruction quality with edge-preserving, noise-reducing effects due to kernel-based modeling.
• Enables adaptive processing—such as varying the number of kernels or quantization strategies for optimized trade-offs between quality and efficiency.
• Supports complex signal manipulation like super-resolution or multi-scale analysis using analytical models.
• The approach is highly modular, suiting a range of hardware and software signal processing applications.

Technical Classifications (CPCs)

Inventors & Applicants

Patent Abstract

An embodiment of the invention relates to a method for processing an input signal (S) and generating an output signal (Sout) based on the input signal (S), said method comprising the steps of: In an analysis stage, extracting a plurality of kernels from the input signal (S), wherein each kernel is described by a parameter vector that is defined by a given number of extracted kernel parameters, and forming the output signal (Sout) based on the extracted kernel parameters.