Deep Learning Voice Extraction for Selective Noise Cancellation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing noise-canceling methods that prevent ambient noise also block nearby voices, making communication difficult in environments where both are needed.
Innovation Solution
A noise canceling method using a deep learning algorithm to extract voice signals from ambient noise and adjust output volume based on probability values, allowing for selective cancellation of ambient noise while preserving nearby voices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If noise canceling method is used to prevent ambient noise, then ambient noise is reduced, but nearby voices are also blocked
Solution Approach 1:
The patent segments the audio signal into different components (ambient noise and nearby voices) and applies different processing strategies to each. The noise canceling device selectively cancels ambient noise while preserving nearby voices by analyzing signal characteristics and applying appropriate filtering methods to different signal components.
Solution Approach 2:
The patent applies different quality characteristics to different parts of the audio signal. Ambient noise regions are heavily filtered while nearby voice regions are preserved with minimal filtering. This is achieved through analyzing signal properties and adaptively adjusting the noise canceling strength based on the local characteristics of each signal segment.
2Measurement precision
If deep learning algorithm is used to extract voice signals, then voice extraction accuracy is improved, but device complexity increases
Solution Approach 1:
The patent employs preliminary action by pre-training deep learning models with extensive training data before deployment. The voice extraction algorithm is pre-trained to recognize various voice patterns and noise conditions, enabling accurate real-time extraction without requiring complex runtime computations. This shifts computational complexity from runtime operation to offline training phase.
Solution Approach 2:
The patent replaces traditional mechanical signal processing methods with deep learning-based artificial intelligence approaches. Instead of using fixed filtering algorithms or manual signal processing techniques, the system uses trained neural networks that automatically learn optimal extraction strategies from data, achieving superior accuracy with more elegant computational solutions.
Data Source
AI summary
Disclosed is a method and device for canceling noise by using a deep learning algorithm. The method includes collecting a noise signal, obtaining a first sound signal, which is obtained by extracting only a voice signal from the collected noise signal, and ‘P’ being a probability value indicating that a human voice signal is included in the collected noise signal, through a deep learning algorithm, and on a basis of a value of the ‘P’, outputting the first sound signal or a second sound signal obtained by converting an overall volume of the collected noise signal. At this time, the second sound signal may be a sound signal, of which a reduction ratio of a volume is converted to be great as the volume corresponds to a great portion, from among the collected noise signal.