Harishchandra (Hari) Dubey
PhD candidate, Electrical Engineering |
|
Advisors: Professor Dr. John H.L. Hansen and Dr. Abhijeet Sangwan |
Bio
Starting January 2016, I am research assistant (PhD candidate) at the Centre for Robust Speech Systems (CRSS) headed by Professor Dr. John H.L. Hansen at the University of Texas at Dallas. My PhD research interests are broadely in Statistical Learning, Estimation and Detection Theory for solving problems in Audio, Speech and Language Processing. Specifically, I am working on Robust Multi-stream Speaker Diarization targeted for studying Peer-Led Team Learning (PLTL) groups that is a special case of small-group conversations. I am also interested in unsupervised/semi-supervised, noise-and reverberation-robust Speech Activity Detection (SAD). The end goal of my research is mining meaningful analytics from audio recordings of small-group conversations such as PLTL.
I received the Bachelor of Technology degree in Electronics and Communication Engineering from Motilal Nehru National Institute of Technology (MNNIT) Allahabad, India in 2012 and the Master of Science degree in Communication and Multimedia Engineering from Friedrich-Alexander University of Erlangen-Nuremberg, Germany in 2015 where I was affiliated with Audio and Acoustic Signal Processing group headed by Professor Dr.-Ing. Walter Kellermann. My master thesis entitled "Distortion Modeling for Robust DNN-based Speech Recognition" proposed a non-linear model for feature enhancement. I was awarded STIBET Stipend, German Academic Exchange Service (DAAD) for Masters Thesis. I enjoyed my wonderful stay in Germany.
I was Visiting Scholar in the department of Electrical, Computer and Biomedical Engineering at the University of Rhode Island (URI), Kingston, USA from April to Dec. 2015 affiliated with Wearable Biosensing Lab headed by Professor Dr.-Ing. Kunal Mankodiya. I worked on wearable Internet-of-Things (IoT) solutions targeted at Voice and Speech Therapy of patients with Parkinson's Disease. I got the URI IP award for this work in Februray 2016. I further explored Fog computing for enhanced speech treatments. I was involved in BigEAR project for audio-based analysis of couples coping with breast cancer.
From 2012 to 2013, I was the Member of Technical Staff at Siemens Ltd., India. I had part-time research appointments at Fraunhofer IIS, International AudioLabs, Erlangen during 2013-2015 in High Definition Video Coding and Spatial Audio Processing departments. I was part-time research associate at Siemens Healthcare Erlangen (now known as Siemens Healthineers) in the department of Angiography and Interventional X-Ray Systems, and Imaging Solutions group during 2013-2015. My work at Siemens Eralngen was broadely in Computed Tomography (CT), motion compensation in surgical images, and contrast-improvement. I was recognized by IP disclosure award at Siemens and contributed several IP blocks in image processing pipeline.
Primary Research Interests
Robust Speaker Diarization and Speech Activity Detection
- Multi-stream Data
- Unsupervised Deep Learning
- Noise-and Reverberation-Robustness
- Discriminative Features
- Statistical Change Detection
- Clustering
Robust Speech Recognition
- Distortion Modeling
- Feature Enhancement
- Adaptation of Deep Neural Network-based Acoustic Model
Secondary Research Interests
- Speech and Voice Therapy (Parkinson's disease)
- Computer-assisted Surgery (Neuro-and-Angiography)
- Wearable and Wireless Health Technologies: Heart Rate & Respiration Rate Estimation
- Fog Computing
Recent Papers
- H. Dubey, L. Kaushik, A. Sangwan, and John H.L. Hansen,"A Speaker Diarization System for Studying Peer-Led Team Learning Groups", in the Proceedings of INTERSPEECH 2016, San Francisco, USA, [PDF], [Bib].
My advisors are Professor Dr. John H.L. Hansen and Abhijeet Sangwan.
For more details on my research projects and education, please visit my
Webpage.
Find me on
Personnel
- Prof. John H.L. Hansen
- Prof. Yang Liu
- Prof. Philip Loizou
- Prof. Abhijeet Sangwan
- Prof. Hynek Boril
- Prof. Oldooz Hazrati
- Dr. Finnian Kelly
- Chengzhu Yu
- Lakshmish Kaushik
- Seyedmahdad (Matt) Mirsamadi
- Abhinav Misra
- Qian Zhang
- Yang ZhengShivesh Ranjan
- Chunlei Zhang
- Fahimeh Bahmaninezhad
- Harishchandra Dubey
- Ahmet Bulut
- Shahram Ghorbani
- Time Sheets