Vikram Nelvoy Rajendiran — accepted AI-conference papers

2025

Mobile-friendly Image de-noising: Hardware Conscious Optimization for Edge Application

Srinivas Soumitri Miriyala, Sowmya Vajrala, Hitesh Kumar, Sravanth Kodavanti, Vikram Nelvoy Rajendiran

ICASSP 2025accepted

Image enhancement is a critical task in computer vision and photography that is often entangled with noise. This renders the traditional Image Signal Processing (ISP) ineffective compared to the advances in deep learning. However, the success of such methods is increasingly associated with the ease…

Cited by 0 Source ↗Scholar ↗

2024

Edge Deployable Distributed Evolutionary Optimization based Calibration method for Neural Quantization

Utsav Tiwari, Srinivas Soumitri Miriyala, Vikram Nelvoy Rajendiran

ICASSP 2024accepted

Accuracy drop in neural quantization is addressed in prior-art through Post Training Quantization (PTQ) schemes such as Percentile and Range-based calibration that remain sensitive to the data distribution. On the other hand, the sophisticated methods that efficiently handle the variability in data…

Cited by 0 Source ↗Scholar ↗

2024

Learning Representations from Explainable and Connectionist Approaches for Visual Question Answering

Aakansha Mishra, Srinivas Soumitri Miriyala, Vikram Nelvoy Rajendiran

ICASSP 2024accepted

Reasoning conditioned on visual and linguistic information has gained immense importance in recent times. The prior art in Visual Question Answering (VQA) has been predominantly connectionist in nature. To resolve the issues of connectionist AI models, Symbolic models were proposed that allowed for…

Cited by 0 Source ↗Scholar ↗

2024

Mixed Precision Neural Quantization with Multi-Objective Bayesian Optimization for on-Device Deployment

Srinivas Soumitri Miriyala, P. K. Suhas, Utsav Tiwari, Vikram Nelvoy Rajendiran

ICASSP 2024accepted

Mixed-precision quantization has emerged as a solution in recent times for accurate inference of Deep Neural Networks on edge. However the prior-art is far from being deployable on embedded devices due to various practical limitations. In this work, a pipeline is designed that a) performs layer-wise…

Cited by 0 Source ↗Scholar ↗

2023

Receptive Field Reliant Zero-Cost Proxies for Neural Architecture Search

Prateek Keserwani, Srinivas Soumitri Miriyala, Vikram Nelvoy Rajendiran, Pradeep N. Shivamurthappa

ICASSP 2023accepted

Neural Architecture Search (NAS) is a fast growing technology for automatic design of deep-learning architectures. NAS includes three stages: search space design, search strategy, and evaluation criterion. Among these, the evaluation of various architectures is very cost-intensive task. In this work…

Cited by 0 Source ↗Scholar ↗