Rupayan Chakraborty
Tata Consultancy Services
13 Papers
36 Citations
Rupayan Chakraborty is an academic researcher from Tata Consultancy Services. The author has contributed to research in topics: Microphone & Robustness (computer science). The author has an hindex of 5, co-authored 13 publications. Previous affiliations of Rupayan Chakraborty include Polytechnic University of Catalonia.
Chat about Author
Papers
Multi-Conditioning and Data Augmentation Using Generative Noise Model for Speech Emotion Recognition in Noisy Conditions
Upasana Tiwari,Meet H. Soni,Rupayan Chakraborty,Ashish Panda,Sunil Kumar Kopparapu +4 more
- 04 May 2020
TL;DR: This paper proposes multi-conditioning and data augmentation using an utterance level parametric Generative noise model, designed to generate noise types which can span the entire noise space in the mel-filterbank energy domain.
49
Deep Encoded Linguistic and Acoustic Cues for Attention Based End to End Speech Emotion Recognition
Swapnil Bhosale,Rupayan Chakraborty,Sunil Kumar Kopparapu +2 more
- 04 May 2020
TL;DR: An End-to-End model with convolutional layers and multi-head self attention mechanism is proposed for Speech Emotion Recognition (SER) task and the linguistic features are found to be effective for the scripted as well as for the combined scenario that reflects more linguistic information in spoken utterances.
26
•Proceedings Article
Detection and Positioning of Overlapped Sounds in a Room Environment.
Rupayan Chakraborty,Climent Nadeu,Taras Butko +2 more
- 01 Jan 2012
TL;DR: Signal-level fusion and likelihood fusion are tried to combine the information from the two pairs of microphones and blind source separation based on the deflation method and null steering beamforming are used for signal separation.
13
Front-End Feature Compensation and Denoising for Noise Robust Speech Emotion Recognition.
Rupayan Chakraborty,Ashish Panda,Meghna Pandharipande,Sonal Joshi,Sunil Kumar Kopparapu +4 more
- 15 Sep 2019
TL;DR: This work implements and compares different frontend robustness techniques for their efficacy in speech emotion recognition, and shows that along with front-end compensation, applying feature selection to non-MFCC highlevel descriptors results in better performance.
13
•Posted Content
Identification of Dementia Using Audio Biomarkers.
TL;DR: This work analyzes the patients audio excerpts from a clinician-participant conversations taken from the Pitt corpus of DementiaBank database, to identify the speech parameters that best distinguish between MCI, AD and healthy (HC) speech.
11