Segmentation to Sound Conversion
Journal Title: IOSR Journals (IOSR Journal of Computer Engineering) - Year 2014, Vol 16, Issue 3
Abstract
Abstract: Our motive, the task of unsupervised topic segmentation of speech data operating over raw acoustic information. In contrast to existing algorithms for topic segmentation of speech, our approach does not require input transcripts. Our method predicts topic changes by analyzing the distribution of reoccurring acoustic patterns in the speech signal corresponding to a single speaker. The algorithm robustly handles noise inherent in acoustic matching by intelligently aggregating information about the similarity profile from multiple local comparisons. Our experiments show that audio-based segmentation compares favorably with transcript based segmentation computed over noisy transcripts. These results demonstrate the desirability of our method for applications where a speech recognizer is not available, or its output has a high word error rate. Also, this paper describes methods for automatically locating points of significant change in music or audio by analyzing local self-similarity. This method can find individual note boundaries or even natural segment boundaries such as verse/c hours or speech/music transitions, even in the absence of cues such as silence. This approach uses the signal to model itself, and thus does not rely on particular acoustic cues nor requires training. Here present a wide variety of applications, including indexing, segmenting, and beat tracking of music and audio. The method works well on a wide variety of audio sources.
Authors and Affiliations
Anindita Chatterjee, , Himadri Nath Moulick
Keywords
Related Articles
- EP ID EP147422
- DOI 10.9790/0661-16354448
- Views 109
- Downloads 0
Congestion Controlled d-Adapt-OR for Adhoc Networks
Abstract: Opportunistic routing scheme is used to route data packet from source to destination. Broadcasting nature of the wireless networks is utilized here to send data packets. A suitable forwarder node is selec...
Performance Evaluation of the Bingo Electronic Voting Protocol
Abstract: Research in e-voting aims at designing usable and secure electronic voting systems. This paperprovides an empirical analysis of the computational performance of a prototype implementation of the Bingoelec...
Handwritten Character Recognition Based on Zoning Using Euler Number for English Alphabets and Numerals
Abstract: Handwritten Character Recognition has been a challenging research domain due to its diverse applicable environment. Handwriting has always been and will possibly continue to be a means of communication. T...
Numerical Analogy of q-Function
Abstract : This paper is a collection of q analogue of various classical methods for finding solutions of algebraic and transcendental equations. It also deals with comparing classical methods with q methods proposed by...
Semantic Similarity Search Model for Obfuscated Plagiarism Detection in Marathi Language using Fuzzy and Naïve Bayes Approaches
Abstract: Plagiarism detection (PD) in natural language texts is an example of NLP applications that is linked with information retrieval (IR) and soft computing (SC) approaches. Obfuscated plagiarism cases contain invis...