Segment Repetition Based on High Amplitude to Enhance a Speech Emotion Recognition

Prayitno, Bagas Adi; Suyanto, Suyanto

doi:10.1016/j.procs.2019.08.234

Cited by 14 publications

(2 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…1 will be recognized in the form of a mel frequency number line which will then be recognized by the classification algorithm. Before producing a good sound feature, an audio signal, a sound signal requires a process that becomes the stage of an MFCC process, such as windowing with a size ranging from 0.25 ms with windowing, the MFCC can produce the MFCC feature converting the audio signal into a measurement parameter generated by the mel spectrum [9] (Fig. 2).…”

Section: Mel Frequency Cepstral Coefisienmentioning

confidence: 99%

Speech Signal Algorithm Conversion from Sasak Language into Sasak Script with CNN and Rule-Based Method

Aranta¹,

Wijaya²,

Bimatoro³

et al. 2022

Proceedings of the First Mandalika International Multi-Conference on Science and Engineering 2022, MIMSE 2022 (Informatics and

View full text Add to dashboard Cite

Sasak speech-language is can't operate in application such as google speech, Alexa assistance, and so on. It makes society need a new development technology who can solve the problem for sasak language. West Nusa Tenggara Province has more than thousand ancient manuscript writen in lontar leaves and papers, that the manuscript spread in Lombok Island and Sumbawa Island. Speech recognition technique for local language is important topic in computer science research, that it can save the culture from the island. This research should be explained how speech processing method working for conversion analog speech signal in Sasak language into Latin text. This study uses the Mel-Frequency Cepstral Coefficient (MFCC) method as a feature extraction method and Convolutional Neural Network (CNN) as a voice classification method into text, and the Rule Base method with UTF-16 which is used to provide rules on Latin letters that will be converted into text. Sasak characters. The algorithm developed is expected to be able to change 50 sound classes into Sasak letters with good accuracy results or above 88%, in changing the voice of the Sasak language into Sasak script by 90.00%.

show abstract

Section: Mel Frequency Cepstral Coefisienmentioning

confidence: 99%

Speech Signal Algorithm Conversion from Sasak Language into Sasak Script with CNN and Rule-Based Method

Aranta¹,

Wijaya²,

Bimatoro³

et al. 2022

Proceedings of the First Mandalika International Multi-Conference on Science and Engineering 2022, MIMSE 2022 (Informatics and

View full text Add to dashboard Cite

show abstract

“…However, the critical regions in speech signals are not as intuitive as images or texts, making it difficult to understand and analyze the attention maps. In SER, [12] showed that the highest amplitude plays an important role in emotion recognition performance because each emotion has a different decibel level. Based on the study, we hypothesized that detecting the largest amplitude and learning the associated representations can improve emotion recognition performance.…”

Section: Introductionmentioning

confidence: 99%

Improving Speech Emotion Recognition Through Focus and Calibration Attention Mechanisms

Kim¹,

An²,

Kim³

2022

Interspeech 2022

View full text Add to dashboard Cite

Attention has become one of the most commonly used mechanisms in deep learning approaches. The attention mechanism can help the system focus more on the feature space's critical regions. For example, high amplitude regions can play an important role for Speech Emotion Recognition (SER). In this paper, we identify misalignments between the attention and the signal amplitude in the existing multi-head self-attention. To improve the attention area, we propose to use a Focus-Attention (FA) mechanism and a novel Calibration-Attention (CA) mechanism in combination with the multi-head self-attention. Through the FA mechanism, the network can detect the largest amplitude part in the segment. By employing the CA mechanism, the network can modulate the information flow by assigning different weights to each attention head and improve the utilization of surrounding contexts. To evaluate the proposed method, experiments are performed with the IEMOCAP and RAVDESS datasets. Experimental results show that the proposed framework significantly outperforms the state-of-the-art approaches on both datasets.

show abstract

Sentiment analysis in non-fixed length audios using a Fully Convolutional Neural Network

García-Ordás

Aláiz‐Moretón

Benítez-Andrades

et al. 2021

Biomedical Signal Processing and Control

View full text Add to dashboard Cite

Segment Repetition Based on High Amplitude to Enhance a Speech Emotion Recognition

Cited by 14 publications

References 6 publications

Speech Signal Algorithm Conversion from Sasak Language into Sasak Script with CNN and Rule-Based Method

Speech Signal Algorithm Conversion from Sasak Language into Sasak Script with CNN and Rule-Based Method

Improving Speech Emotion Recognition Through Focus and Calibration Attention Mechanisms

Sentiment analysis in non-fixed length audios using a Fully Convolutional Neural Network

Contact Info

Product

Resources

About