A
Anurag Mittal
Researcher at Indian Institute of Technology Madras
Publications - 97
Citations - 4413
Anurag Mittal is an academic researcher from Indian Institute of Technology Madras. The author has contributed to research in topics: Object detection & Pose. The author has an hindex of 31, co-authored 97 publications receiving 3961 citations. Previous affiliations of Anurag Mittal include Cornell University & Princeton University.
Papers
More filters
Posted Content
Domain Adaptive Knowledge Distillation for Driving Scene Semantic Segmentation
TL;DR: This paper presents a novel approach to learn domain adaptive knowledge in models with limited memory, thus bestowing the model with the ability to deal with these issues in a comprehensive manner and introduces a novel cross entropy loss that leverages pseudo labels from the teacher.
Proceedings ArticleDOI
Efficient Space-time Video Super Resolution using Low-Resolution Flow and Mask Upsampling
TL;DR: In this paper, an efficient solution for Space-time Super-Resolution, aiming to generate high-resolution Slow-motion videos from Low Resolution and Low Frame rate videos, is proposed.
Proceedings ArticleDOI
Domain Adaptive Knowledge Distillation for Driving Scene Semantic Segmentation
TL;DR: In this article, a multi-level distillation strategy is proposed to effectively distil knowledge at different levels, and a novel cross entropy loss is introduced to leverage pseudo labels from the teacher.
Journal ArticleDOI
Linguistically-aware attention for reducing the semantic gap in vision-language tasks
TL;DR: An attention mechanism - Linguistically-aware Attention (LAT) - that leverages object attributes obtained from generic object detectors along with pre-trained language models to reduce this semantic gap between the modalities is proposed.
Proceedings ArticleDOI
Adversarial Joint-Distribution Learning for Novel Class Sketch-Based Image Retrieval
TL;DR: A new framework for ZS-SBIR is proposed that models joint distribution between the sketch and image domain using a generative adversarial network and helps to synthesize the novel class image features using sketch features.