Pranay Manocha
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Pranay Manocha has a total of 20 co-authors. Here are the top co-authors by number of shared publications:
- Buye Xu
Meta Reality Labs Research
2 shared publications Anjali Menon
1 shared publication- Ankit Shah
Accenture, USA
1 shared publication Anurag Kumar
Language Technologies Institute, Carnegie Mellon University, Pittsburgh, United States
1 shared publication- Anurag Kumar
ECE Department and the ESE Department, Indian Institute of Science Bangalore (IISc), Bengaluru, India
1 shared publication - Arun Manohar Gurav
Regional Ayurveda Research Institute, Nehru Garden, Pune 411 038, Maharashtra, India
1 shared publication
Latest Publications
- Top Cited PapersOpen Access
- 4 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Torchaudio-Squim: Reference-Less Speech Quality and Intelligibility Measures in Torchaudio
Measuring quality and intelligibility of a speech signal is usually a critical step in development of speech processing systems. To enable this, a variety of metrics to measure quality and intelligibility under different assumptions have been developed. Through this paper, we introduce tools and a set of models to estimate such known metrics using deep neural networks. These models are made available in the well-established TorchAudio library, the core audio and speech processing library within the PyTorch deep learning framework. We refer to it as TorchAudio-Squim, TorchAudio-Speech QUality and Intelligibility Measures. More specifically, in the current version of TorchAudio-squim, we establish and release models for estimating PESQ, STOI and SI-SDR among objective metrics and MOS among subjective metrics. We develop a novel approach for objective metric estimation and use a recently developed approach for subjective metric estimation. These models operate in a "referenceless" manner, that is they do not require the corresponding clean speech as reference for speech assessment. Given the unavailability of clean speech and the effortful process of subjective evaluation in real-world situations, such easy-to-use tools would greatly benefit speech processing research and development.
- 4 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Top Cited PapersOpen AccessTorchaudio-Squim: Reference-Less Speech Quality and Intelligibility Measures in Torchaudio
Measuring quality and intelligibility of a speech signal is usually a critical step in development of speech processing systems. To enable this, a variety of metrics to measure quality and intelligibility under different assumptions have been developed. Through this paper, we introduce tools and a set of models to estimate such known metrics using deep neural networks. These models are made available in the well-established TorchAudio library, the core audio and speech processing library within the PyTorch deep learning framework. We refer to it as TorchAudio-Squim, TorchAudio-Speech QUality and Intelligibility Measures. More specifically, in the current version of TorchAudio-squim, we establish and release models for estimating PESQ, STOI and SI-SDR among objective metrics and MOS among subjective metrics. We develop a novel approach for objective metric estimation and use a recently developed approach for subjective metric estimation. These models operate in a "referenceless" manner, that is they do not require the corresponding clean speech as reference for speech assessment. Given the unavailability of clean speech and the effortful process of subjective evaluation in real-world situations, such easy-to-use tools would greatly benefit speech processing research and development.
- Open Access
SAQAM: Spatial Audio Quality Assessment Metric
Audio quality assessment is critical for assessing the perceptual realism of sounds. However, the time and expense of obtaining ''gold standard'' human judgments limit the availability of such data. For AR&VR, good perceived sound quality and localizability of sources are among the key elements to ensure complete immersion of the user. Our work introduces SAQAM which uses a multi-task learning framework to assess listening quality (LQ) and spatialization quality (SQ) between any given pair of binaural signals without using any subjective data. We model LQ by training on a simulated dataset of triplet human judgments, and SQ by utilizing activation-level distances from networks trained for direction of arrival (DOA) estimation. We show that SAQAM correlates well with human responses across four diverse datasets. Since it is a deep network, the metric is differentiable, making it suitable as a loss function for other tasks. For example, simply replacing an existing loss with our metric yields improvement in a speech-enhancement network.
Open AccessSAQAM: Spatial Audio Quality Assessment Metric
Audio quality assessment is critical for assessing the perceptual realism of sounds. However, the time and expense of obtaining ''gold standard'' human judgments limit the availability of such data. For AR&VR, good perceived sound quality and localizability of sources are among the key elements to ensure complete immersion of the user. Our work introduces SAQAM which uses a multi-task learning framework to assess listening quality (LQ) and spatialization quality (SQ) between any given pair of binaural signals without using any subjective data. We model LQ by training on a simulated dataset of triplet human judgments, and SQ by utilizing activation-level distances from networks trained for direction of arrival (DOA) estimation. We show that SAQAM correlates well with human responses across four diverse datasets. Since it is a deep network, the metric is differentiable, making it suitable as a loss function for other tasks. For example, simply replacing an existing loss with our metric yields improvement in a speech-enhancement network.
- 1 April 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 3136-3140
Content-Based Representations of Audio Using Siamese Neural Networks
In this paper, we focus on the problem of content-based retrieval for audio, which aims to retrieve all semantically similar audio recordings for a given audio clip query. This problem is similar to the problem of query by example of audio, which aims to retrieve media samples from a database, which are similar to the user-provided example. We propose a novel approach which encodes the audio into a vector representation using Siamese Neural Networks. The goal is to obtain an encoding similar for files belonging to the same audio class, thus allowing retrieval of semantically similar audio. Using simple similarity measures such as those based on simple euclidean distance and cosine similarity we show that these representations can be very effectively used for retrieving recordings similar in audio content.
- 1 April 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 3136-3140
Content-Based Representations of Audio Using Siamese Neural Networks
In this paper, we focus on the problem of content-based retrieval for audio, which aims to retrieve all semantically similar audio recordings for a given audio clip query. This problem is similar to the problem of query by example of audio, which aims to retrieve media samples from a database, which are similar to the user-provided example. We propose a novel approach which encodes the audio into a vector representation using Siamese Neural Networks. The goal is to obtain an encoding similar for files belonging to the same audio class, thus allowing retrieval of semantically similar audio. Using simple similarity measures such as those based on simple euclidean distance and cosine similarity we show that these representations can be very effectively used for retrieving recordings similar in audio content.
- 1 February 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 712-716
Tumor Segmentation and Gradation for MR Brain Images
MRI is a non-invasive technology that is currently used for analyzing brain tumors, which can only conclusively be graded by invasive tissue extraction. Techniques like perfusion used for analysis of tumors are time taking, therefore, there is increased focus towards automated computational techniques to segment and grade tumors using standard MR sequences like T1 and T2 weighted images. These are intensity based images and are not enough to grade tumors, thus, are not used in clinical practice. This work automates tumor segmentation without the use of templates and training sets, by using vertical symmetry comparison. Once the tumor is segmented, fuzzy c means clustering is used on T1 and T2 maps, that are generated by using three intensity based sequences namely T1, T2 and proton density (PD) fat saturated images, for tumor grading. The clustered sections are compared with cerebral blood volume (CBV) images that are obtained by perfusion parameters to find correlation. 15 patients, 5 each from 4 th , 3 rd and 2 nd grade tumor categories, were tested. Some correlation is seen between T1, T2 map values of tumor region and CBV values showing that this area can be explored further for tumor grading.
- 1 February 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 712-716
Tumor Segmentation and Gradation for MR Brain Images
MRI is a non-invasive technology that is currently used for analyzing brain tumors, which can only conclusively be graded by invasive tissue extraction. Techniques like perfusion used for analysis of tumors are time taking, therefore, there is increased focus towards automated computational techniques to segment and grade tumors using standard MR sequences like T1 and T2 weighted images. These are intensity based images and are not enough to grade tumors, thus, are not used in clinical practice. This work automates tumor segmentation without the use of templates and training sets, by using vertical symmetry comparison. Once the tumor is segmented, fuzzy c means clustering is used on T1 and T2 maps, that are generated by using three intensity based sequences namely T1, T2 and proton density (PD) fat saturated images, for tumor grading. The clustered sections are compared with cerebral blood volume (CBV) images that are obtained by perfusion parameters to find correlation. 15 patients, 5 each from 4 th , 3 rd and 2 nd grade tumor categories, were tested. Some correlation is seen between T1, T2 map values of tumor region and CBV values showing that this area can be explored further for tumor grading.