Audio Identification by Sampling Sub-fingerprints and Counting Matches

Yao, Shanshan; Niu, Baoning; Liu, Jianquan

doi:10.1109/tmm.2017.2723846

Cited by 8 publications

(15 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In terms of robustness analysis, this experiment uses the feature dimension reduction algorithms of Section 3.2.2 and Section 3.2.3 to extract MFCC feature and LPCC feature of the original speech to construct audio fingerprint, this paper compares the audio fingerprint based on the combined features with the audio fingerprints of two features and the existing audio fingerprint methods of [6,7,22,23,25,27], where [6,7] is an improved method based on Shazam fingerprint, [22,23,25,27] is an improved method based on Philips fingerprint. As shown in Table 1, the combined feature is more robust than MFCC and LPCC that under the same feature dimension reduction process.…”

Section: Robustness and Retrieval Performance Analysismentioning

confidence: 99%

“…As shown in Table 1, the combined feature is more robust than MFCC and LPCC that under the same feature dimension reduction process. The proposed method has good retrieval performance under different CPOs, compared with the audio fingerprint algorithm based on short speech segments in [6,7,22,23,25] under several CPOs, the proposed method can get similar or even better recall rate and precision rate, and compared with the audio fingerprint algorithm based on long speech segments that is robust to noise proposed in [27], recall rate and precision rate of the proposed method in terms of background noise and narrowband Gaussian noise are higher than [27].…”

Section: Robustness and Retrieval Performance Analysismentioning

confidence: 99%

“…It can be seen from Table 2 that after feature dimension reduction based on information entropy and energy, combined features have better retrieval performance than MFCC and LPCC that under the same feature dimension reduction processing. In addition, the proposed method can ensure high retrieval performance after MP3 compression, resample, background noise, factory noise and narrowband Gaussian noise, and the retrieval performance is higher than [6,7,22,23,25,27].…”

Section: Robustness and Retrieval Performance Analysismentioning

confidence: 99%

“…Wang et al [21] proposed an entropy-based audio fingerprint retrieval technology, which uses the maximum common string, edit distance, and DTW algorithm to achieve retrieval in the retrieval phase. Yao et al [22] significantly increased the retrieval speed of audio fingerprints through sampling and counting methods during the retrieval phase, using the inverted index reduced the storage space of audio fingerprints, which has higher recall rate and precision rate. Sun et al [23] optimized the hash table of audio fingerprints through the Fibonacci sequence and adjusted the length of the hash table through the right shift operation, which effectively reduced memory consumption and improved system retrieval efficiency.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Audio Fingerprint Retrieval Method Based on Feature Dimension Reduction and Feature Combination

2021

KSII TIIS

View full text Add to dashboard Cite

In order to solve the problems of the existing audio fingerprint method when extracting audio fingerprints from long speech segments, such as too large fingerprint dimension, poor robustness, and low retrieval accuracy and efficiency, a robust audio fingerprint retrieval method based on feature dimension reduction and feature combination is proposed. Firstly, the Mel-frequency cepstral coefficient (MFCC) and linear prediction cepstrum coefficient (LPCC) of the original speech are extracted respectively, and the MFCC feature matrix and LPCC feature matrix are combined. Secondly, the feature dimension reduction method based on information entropy is used for column dimension reduction, and the feature matrix after dimension reduction is used for row dimension reduction based on energy feature dimension reduction method. Finally, the audio fingerprint is constructed by using the feature combination matrix after dimension reduction. When speech's user retrieval, the normalized Hamming distance algorithm is used for matching retrieval. Experiment results show that the proposed method has smaller audio fingerprint dimension and better robustness for long speech segments, and has higher retrieval efficiency while maintaining a higher recall rate and precision rate.

show abstract

Section: Robustness and Retrieval Performance Analysismentioning

confidence: 99%

Section: Robustness and Retrieval Performance Analysismentioning

confidence: 99%

Section: Robustness and Retrieval Performance Analysismentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Audio Fingerprint Retrieval Method Based on Feature Dimension Reduction and Feature Combination

2021

KSII TIIS

View full text Add to dashboard Cite

show abstract

“…A long frame length resulted in the spectrogram having low time resolution, which makes the representation insensitive to time variations [15]. In addition, large overlap is an advantage when dealing with using short query music against the long original signal [16][17].…”

Section: Spectrogram Constructionmentioning

confidence: 99%

Genre Classification of Traditional Malay Music Using Spectrogram Correlation

Samad¹,

Huddin

2018

IJET

View full text Add to dashboard Cite

A method to classify the genre of traditional Malay music using spectrogram correlation is described. The method can be divided into three distinct parts consisting of spectrogram construction that retains the most salient feature of the music, template construction that takes into account the variations in music within a genre as well as the music progresses, and template matching based on spectrogram image cross-correlation with unconstrained minimum average correlation energy filters. Experiments conducted with seven genres of traditional Malay music show that the recognition accuracy is dependent on the number of segments used to construct the filter templates, which in turn is related to the length of music segment used. Despite using a small dataset, an average recognition rate of 61.8 percent was obtained for music segments lasting 180 seconds using six relatively short excerpts.

show abstract

Accuracy comparisons of fingerprint based song recognition approaches using very high granularity

Serrano

Scarpa

2023

Multimed Tools Appl

View full text Add to dashboard Cite

Music and song recognition is an activity of wide interest for researchers and companies due to the intrinsic challenges and the possible economical profits it can give. Despite basic algorithms about song recognition are simple in principle, it is quite difficult to obtain an efficient and robust approach able to generate an effective algorithm for identifying short piece of audio on the fly. In this paper, we compare the results obtained using a new algorithm we recently proposed against several baseline approaches in terms of accuracy when very short pieces of audio are processed. Experimental results, performed using both a subset of the MTG-Jamendo dataset and a proprietary audio corpus containing 7000 songs, show our approach outperform the others in particular for excerpts of audio shorter than 3 s .

show abstract

Audio Identification by Sampling Sub-fingerprints and Counting Matches

Cited by 8 publications

References 13 publications

Audio Fingerprint Retrieval Method Based on Feature Dimension Reduction and Feature Combination

Audio Fingerprint Retrieval Method Based on Feature Dimension Reduction and Feature Combination

Genre Classification of Traditional Malay Music Using Spectrogram Correlation

Accuracy comparisons of fingerprint based song recognition approaches using very high granularity

Contact Info

Product

Resources

About