Method, device and equipment for quickly judging and compressing FFT (Fast Fourier Transform) features and storage medium
Through the FFT feature rapid judgment and compression method, the problems of inefficient, insufficient accuracy and high resource consumption of specific sound or transaction judgment methods in the prior art are solved, and efficient and accurate feature judgment and data compression are achieved.
CN119724203AInactive Publication Date: 2025-03-28深圳市友杰智新科技有限公司
9 Cites 0 Cited by
Patent Information
- Application Number
- CN202510247985.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-04
- Publication Date
- 2025-03-28
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Technical Problem
The existing method of judging specific sounds or transactions is inefficient, has insufficient accuracy, and has a large resource consumption in the computing process.
Method used
The FFT feature rapid judgment and compression method is used to identify the preset sound frequency or transaction feature range, perform fast Fourier transformation, find feature locations, perform mathematical calculations, determine whether the feature value exceeds the threshold, and perform compression buffering or reasoning based on the judgment results.
Benefits of technology
Significantly reduces computing time and resource consumption, improves processing speed and accuracy, achieves high compression ratios, and reduces development and maintenance costs.
✦ Generated by Eureka AI based on patent content.
Abstract
The invention relates to the technical field of voice recognition, in particular to an FFT feature rapid judgment and compression method and device, equipment and a storage medium, and the method comprises the steps: recognizing a preset sound frequency or transaction feature range, and obtaining a first specified sound frequency or a first specified feature; performing fast Fourier transform on the input original data to obtain corresponding frequency domain data; searching the position of the first specified sound frequency or the first specified feature in the frequency domain data; performing mathematical calculation based on the search result to obtain a feature value; judging whether the feature value exceeds a first preset threshold value or not; and carrying out compression caching or reasoning on the characteristic value based on a judgment result. According to the method, the transaction standard can be quickly and accurately judged through one-time FFT calculation, the calculation time consumption is reduced compared with a complex AI neural network, and meanwhile higher accuracy and lower storage requirements are provided compared with direct processing of original data.
Need to check novelty before this filing date? Find Prior Art