2025 IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2025) is a flagship conference on speech and audio processing, which was held in Hyderabad from 6 to 11 April. IIIT Hyderabad participated very actively in ICASSP 2025 in various capacities. Faculty, students and staff from IIIT Hyderabad have played their fair share in the successful organization of the conference.
The following are the highlights of IIIT Hyderabad @ ICASSP:
- IIIT Hyderabad was a sponsor of ICASSP. A sponsorship booth was given to IIIT, where posters related to some centers of interest to ICASSP and smart city living lab were showcased. Prof. Santosh coordinated the setup of the IIIT booth at the conference. The sponsorship benefit itself was offered because IIITH provided eduroam support to the entire conference by installing a server at the venue. Sriranjani and Prakash, along with Prof. Sachin, held multiple discussions with the organizing committee and HICC before the conference. Madhubabu from the IT team gave in-person support at the conference venue. Hundreds of attendees joined the network through the eduroam server.
- Prof. Jayanthi served on the awards committee. The awards committee was responsible for selecting the best papers in the conference. She also gave an invited talk in the NeuroDA workshop and was a panelist in the PROGRESS workshop.
- Prof. Sachin co-chaired the ICASSP Show and Tell Demos session. The ICASSP 2025 Show and Tell emphasized the demonstration of novel and interactive signal processing innovations. A total of 47 proposals were submitted from around the world.The Show and Tell sessions were a big success, drawing large crowds of curious attendees. People actively engaged with the presenters, asking questions and having lively discussions about the demos.
- Prof. Lalitha served as the local organization committee chair of the conference. She (along with the other co-chairs) was responsible for most of the local arrangements and contributed heavily along with the rest of the organizing committee.
- Prof. Anil Vuppala (along with his students) entirely took care of the poster printing activity. This activity was very crucial to the conference since for around 1300 posters, the authors could not come to the conference. Based on the available posters and video presentations in the conference portal, around 1200 posters were printed along with the video QR code. Prof. Anil and his students have taken care of this humongous activity, all by themselves.
- Student volunteers: Around 30 students from IIITH (primarily from SPCRC, LTRC and CVIT) very actively volunteered for the organization of ICASSP. They had volunteered for more than half of the entire poster sessions (this is around 150+ in number). As part of volunteering, the students would put up posters in the poster area and ensure the smooth organization of the poster sessions. IIIT students were the most active volunteers in the entire conference. They also helped with setting up videos, helped the general chairs with ICASSP50 Pavilion and many other tasks.
- Prof. Bapi Raju organized the Neuro Data Analytics workshop at the conference and Prof. Vinoo Alluri co- organized the workshop on Indian Music Analysis and Generative Applications.
List of papers and posters presented :
- Typical vs. Atypical Disfluency Classification: Introducing the IIITH-TISA Corpus and Temporal Context-Based Feature Representations – Priyanka Kommagouni1, Vamshiraghusimha Narasinga, Purva Barche, Sai Akarsh C, Anil Vuppala
- Enhancing Stutter Detection using Long-Term Average Spectrum Values – Vamshiraghusimha Narasinga, Priyanka Kommagouni, Sridhar Vanga, Kowshik Siva Sai Motepalli, Sai Akarsh C, Purva Barche, Anil Vuppala
- A Multi-modal Approach to Dysarthria Detection and Severity Assessment Using Speech and Text Information – Anuprabha M; Krishna Gurugubelli; Samsung Research & Development Institute-Bengaluru; Kesavaraj V and Anil Kumar Vuppala
- Evaluating the Impact of Discriminative and Generative E2E Speech Enhancement Models on Syllable Stress Preservation – Rangavajjala Sankara Bharadwaj, Jhansi Mallela, Sai Harshitha Aluru, Chiranjeevi Yarra
- Post-Net2.0: An adaptive weighted loss function driven by linguistic constraint for automatic syllable stress detection – Sai Harshitha Aluru, Jhansi Mallela, Chiranjeevi Yarra
- MRI2Speech: Speech Synthesis from Articulatory Movements Recorded by Real-time MRI – Neil Shah, Ayan Kashyap, Shirish Karande, Vineet Gandhi
- Prompt-to-Correct: Automated Test-Time Pronunciation Correction with Voice Prompts – Ayan Kashyap, Neil Shah, Vineet Gandhi
- Advancing NAM-to-Speech Conversion with Novel Methods and MultiNAM Dataset – Neil Shah, Shirish Karande, Vineet Gandhi
- Minimalistic Video Saliency Prediction via Efficient Decoder & Temporal Action Cues – Rohit Giramji, Siddharth Jain, Bhav Beri, Sarthak Bansal, Vineet Gandhi
- Sequential DOA Trajectory Estimation using Deep Complex Network and Residue Signals – Shreyas Jaiswal, Peter Gerstoft, Santosh Nannuru
List of Show and Tell Demos presented:
- Waveform-Wizard: A free tool for Spectro-Temporal visualization of Speech – Anil Vuppala, Anirudh Bocha, Abhinav Reddy, Sasidhar Chavali, Saikiran Survi, Srivatsa Losetti, Sree Harsha, Sumanth Vuddagiri, Harshwardhan, Srihari Bandarupalli, Vamshiraghusimha Narasinga, Purva Barche
- Signals and Systems Virtual Lab – Santosh Nannuru, Sasanka G R S, Ruchi Pandey, Krish Pandya
- Building a Field Area Network using Wi-SUN in the License-free sub-GHz Band – Anuradha Vattem, Kartik Gharde
Lecture Demonstrations on Indian Dance and music as well as Yoga sessions: Dr. Jayachandran and team presented a lecture demonstration on Indian dance forms.This clearly demonstrated the similarities and differences between the various dance forms and also the rich diversity of the country. The lecture demonstration went on for about an hour and around 30 participants attended the event. Dr. Saroja and team presented a lecture demonstration on Indian music. The presentation was accompanied by the instrumental support of Tanpura, Mridangam, Ghatam, Tabla, Violin and Flute serving as a few examples to demonstrate the sonic landscape of Indian music. The duration of the lecture demonstration was an hour and a half. Around 40 participants, including many foreign nationals, attended the event. Prof. Suresh Purini conducted Heartfulness Yoga sessions on all 4 days in ICASSP. These sessions included gentle yoga poses, breathing exercises, and meditation. Around 20 conference attendees participated in the yoga sessions.
Networking Event at IIIT Hyderabad: A networking event was organized just before the start of the main ICASSP conference. Around 10 invitees visited IIIT-H including Prof. Mos Kaveh and Prof. Giannakis. Several other visitors came to IIIT Hyderabad in that week including Prof. Pramod Varshney, Prof. Rama Chellappa, Prof. Dinesh Manocha and others.
April 2025