Preprint2026Under review
@misc{singh2026mitigating,
title = {{Mitigating Distribution Shift for Adversarial Purification}},
author = {Himanshu Singh and A V Subramanyam},
note = {Under review},
year = {2026}
}Preprint2026Under review
* Equal contribution
@misc{rajput2026learning,
title = {{Learning Adversarial Manifold for Jailbreaking at Scale}},
author = {Shivank Rajput and Himanshu Singh and A V Subramanyam},
note = {Under review},
year = {2026}
}arXiv2026
@misc{chatterjee2026unified,
title = {{Unified Multi-Dataset Training for TBPS}},
author = {Nilanjana Chatterjee and Sidharatha Garg and A V Subramanyam and Brejesh Lall},
eprint = {2601.14978},
archivePrefix = {arXiv},
year = {2026}
}arXiv2026
@misc{mishra2026scene,
title = {{Scene-Consistent Illumination Transfer for Inserted Advertising Graphics}},
author = {Rameshwar Mishra and Bishshoy Das and A. V. Subramanyam and Guan-Ming Su},
eprint = {2609.37951},
archivePrefix = {arXiv},
year = {2026}
}arXiv2026
@misc{jayasundara2026corrupting,
title = {{Corrupting Attention: Evasion-Based Adversarial Attacks on Encoder Attention in Detection Transformers}},
author = {Ridma Jayasundara and Shaheer Mohamed and Tharindu Fernando and Harshala Gammulle and Basura Fernando and Sanka Rasnayake and A V Subramanyam and Sridha Sridharan and Clinton Fookes},
eprint = {2608.06674},
archivePrefix = {arXiv},
year = {2026}
}arXiv2025
@misc{singh2025novel,
title = {{Novel View Synthesis using DDIM Inversion}},
author = {Sehajdeep Singh and A V Subramanyam and Aditya Gupta and Sahil Gupta},
eprint = {2508.10688},
archivePrefix = {arXiv},
year = {2025}
}arXiv2025
@misc{modi2025boosting,
title = {{Boosting Weak Positives for Text Based Person Search}},
author = {Akshay Modi and Ashhar Aziz and Nilanjana Chatterjee and A V Subramanyam},
eprint = {2501.17586},
archivePrefix = {arXiv},
year = {2025}
}arXiv2025
@misc{bhagat2025conditional,
title = {{Conditional Consistency Guided Image Translation and Enhancement}},
author = {Amil Bhagat and Milind Jain and A. V. Subramanyam},
eprint = {2501.01223},
archivePrefix = {arXiv},
year = {2025}
}NeurIPS2026
Conference on Neural Information Processing Systems (NeurIPS) · to appear
Proposes an inversion-free, text-driven image editing method that steers flow-model editing trajectories using entropic optimal transport between source and target latent particle clouds, improving structure preservation and semantic alignment.
@inproceedings{mishra2026otedit,
title = {{OTEdit: Entropic Transport Corrected Trajectory for Inversion-Free Flow-Based Image Editing}},
author = {Rameshwar Mishra and Akshay Modi and A V Subramanyam},
booktitle = {Conference on Neural Information Processing Systems (NeurIPS)},
month = dec,
year = {2026}
}BMVC2026
British Machine Vision Conference (BMVC) · to appear
@inproceedings{mishra2026refineflow,
title = {{RefineFlow: Estimating the Editing Velocity with Geometric, Trajectory, and Locality Priors for Inversion Free Flow Editing}},
author = {Rameshwar Mishra and A V Subramanyam},
booktitle = {British Machine Vision Conference (BMVC)},
month = nov,
year = {2026}
}EMNLP2026
Findings of the Association for Computational Linguistics: EMNLP 2026 · to appear
Proposes a gradient-based subspace intervention that finds and suppresses hidden toxic patterns in LLM representations, reducing toxicity of state-of-the-art detoxification systems by 8-20% while keeping fluency.
Large Language Models (LLMs) are powerful text generators, yet they can produce toxic or harmful content even when given seemingly harmless prompts. This presents a serious safety challenge and can cause real-world harm. Toxicity is often subtle and context-dependent, making it difficult to detect at the token level or through coarse sentence-level signals. Moreover, efforts to mitigate toxicity often face a trade-off between safety and the coherence, or fluency, of the generated text. In this work, we present a targeted subspace intervention strategy based on gradient signals for identifying and suppressing hidden toxic patterns from underlying model representations, while preserving the overall ability to generate safe, fluent content. On the RealToxicityPrompts dataset, our method achieves strong mitigation performance compared to existing baselines while imposing minimal impact on inference complexity. Across LLMs, our approach reduces the toxicity of state-of-the-art detoxification systems by 8-20%, while maintaining comparable fluency. We further validate cross-benchmark generality on ToxiGen, where the same gains hold under adversarial, implicitly toxic prompts. Through extensive quantitative and qualitative analyses, we show that our approach achieves effective toxicity reduction without impairing generative performance, consistently outperforming existing baselines.
@inproceedings{singh2026do,
title = {{Do Prompts Guarantee Safety? Mitigating Toxicity from LLM Generations through Gradient Based Subspace Intervention}},
author = {Himanshu Singh and Ziwei Xu and A V Subramanyam and Mohan Kankanhalli},
booktitle = {Findings of the Association for Computational Linguistics: EMNLP 2026},
eprint = {2602.06623},
archivePrefix = {arXiv},
doi = {10.48550/arXiv.2602.06623},
month = oct,
year = {2026}
}ECCV2026
European Conference on Computer Vision (ECCV), LNCS Part XXV, pp. 366-382
RefDiT is a reference-guided image generation framework that decomposes identifier tokens into attribute-level signals and trains LoRA blocks of a diffusion transformer to transfer local attributes from reference images.
@inproceedings{mishra2026refdit,
title = {{RefDiT: Local Attribute Guidance in Reference-Based Image Generation}},
author = {Rameshwar Mishra and Srikrishna Karanam and A V Subramanyam},
booktitle = {European Conference on Computer Vision (ECCV), LNCS Part XXV},
eprint = {2609.04976},
archivePrefix = {arXiv},
doi = {10.1007/978-3-032-37461-5_20},
month = sep,
year = {2026}
}IEEE TAI2026
IEEE Transactions on Artificial Intelligence, vol. 7, no. 9, pp. 5287-5295
Proposes an adversarial training method that improves robustness by finding each adversarial sample's nearest neighbours from other classes and removing its feature projections onto them, reducing inter-class feature overlap.
Deep neural networks have exhibited impressive performance in image classification tasks but remain vulnerable to adversarial examples. Standard adversarial training enhances robustness but typically fails to explicitly address inter-class feature overlap, a significant contributor to adversarial susceptibility. In this work, we introduce a novel adversarial training framework that actively mitigates inter-class proximity by projecting out inter-class dependencies from adversarial and clean samples in the feature space. Specifically, our approach first identifies the nearest inter-class neighbors for each adversarial sample and subsequently removes projections onto these neighbors to enforce stronger feature separability. Theoretically, we demonstrate that our proposed logits correction reduces the Lipschitz constant of neural networks, thereby lowering the Rademacher complexity, which directly contributes to improved generalization and robustness. Extensive experiments across standard benchmarks including CIFAR-10, CIFAR-100, and SVHN show that our method demonstrates strong performance that is competitive with leading adversarial training techniques, highlighting significant achievements in both robust and clean accuracy. Our findings reveal the importance of addressing inter-class feature proximity explicitly to bolster adversarial robustness in DNNs.
@article{singh2026nearest,
title = {{Nearest Neighbor Projection Removal Adversarial Training}},
author = {Himanshu Singh and A. V. Subramanyam and Shivank Rajput and Mohan Kankanhalli},
journal = {IEEE Transactions on Artificial Intelligence},
eprint = {2509.07673},
archivePrefix = {arXiv},
doi = {10.1109/TAI.2026.3681719},
month = sep,
year = {2026}
}ICMEW2026
IEEE International Conference on Multimedia and Expo Workshops (ICMEW), CMI2IT Workshop, Bangkok, Thailand · to appear
Presents AD-Relight, a multi-stage, training-free framework that adapts a diffusion-based relighting model at test time to realistically relight custom ad banners inserted into scenes.
@inproceedings{mishra2026ad,
title = {{AD-Relight: Training-Free Banner Relighting via Illumination Translation with Diffusion Priors}},
author = {Rameshwar Mishra and A V Subramanyam},
booktitle = {IEEE International Conference on Multimedia and Expo Workshops (ICMEW), CMI2IT Workshop, Bangkok, Thailand},
eprint = {2604.24407},
archivePrefix = {arXiv},
month = jul,
year = {2026}
}ICASSP2026
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
Studies neural network training dynamics through Lie brackets of loss-gradient vector fields: the brackets vanish in the NTK (lazy) regime, and non-zero brackets mark feature learning.
@inproceedings{subramanyam2026lie,
title = {{Lie Bracket Geometry of Feature Learning in Neural Networks}},
author = {A V Subramanyam},
booktitle = {IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
doi = {10.1109/icassp55912.2026.11462529},
month = may,
year = {2026}
}ICME2025
IEEE International Conference on Multimedia and Expo (ICME), pp. 1-6
Proposes the Beckman barycenter, an efficiently computable optimal-transport barycenter, which is used to fine-tune adversarially trained networks and improves their robustness to AutoAttack on CIFAR and Tiny ImageNet.
@inproceedings{subramanyam2025beckman,
title = {{Beckman Adversarial Defense}},
author = {A. V. Subramanyam},
booktitle = {IEEE International Conference on Multimedia and Expo (ICME)},
eprint = {2301.01495},
archivePrefix = {arXiv},
doi = {10.1109/ICME59968.2025.11209355},
year = {2025}
}ICASSP2025
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 1-5
Proposes KAMIM, which weights image patches using keypoint features during SimMIM's masked reconstruction, improving linear probing and fine-tuning accuracy of vision transformers on ImageNet-1K.
@inproceedings{krishna2025keypoint,
title = {{Keypoint Aware Masked Image Modelling}},
author = {Madhava Krishna and A. V. Subramanyam},
booktitle = {IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
eprint = {2407.13873},
archivePrefix = {arXiv},
doi = {10.1109/ICASSP49660.2025.10889887},
year = {2025}
}ICASSP2025
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Hyderabad, India, pp. 1-5
Introduces a memory-efficient patch-based framework that combines global context with local patch information to train on ultra-high-resolution images, running even on resource-constrained devices like Jetson Nano.
@inproceedings{subramanyam2025efficient,
title = {{Efficient Localized Perception for Resource-Constrained Vision Systems}},
author = {A. V. Subramanyam and Niyati Singal and Vinay K. Verma},
booktitle = {IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Hyderabad, India},
eprint = {2405.07166},
archivePrefix = {arXiv},
doi = {10.1109/ICASSP49660.2025.10888483},
year = {2025}
}BMVC2024
The British Machine Vision Conference (BMVC)
@inproceedings{mishra2024layout,
title = {{Layout Free Scene Graph to Image Generation}},
author = {Rameshwar Mishra and A V Subramanyam},
booktitle = {The British Machine Vision Conference (BMVC)},
doi = {10.48550/arXiv.2401.14111},
year = {2024}
}ICASSP2024
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
Adversarial purification using generative models demonstrates strong adversarial defense performance. These methods are classifier and attack-agnostic, making them versatile but often computationally intensive. Recent strides in diffusion and score networks have improved image generation and, by extension, adversarial purification. Another highly efficient class of adversarial defense methods known as adversarial training requires specific knowledge of attack vectors, forcing them to be trained extensively on adversarial examples. To overcome these limitations, we introduce a new framework, namely Language Guided Adversarial Purification (LGAP), utilizing pre-trained diffusion models and caption generators to defend against adversarial attacks. Given an input image, our method first generates a caption, which is then used to guide the adversarial purification process through a diffusion network. Our approach has been evaluated against strong adversarial attacks, proving its effectiveness in enhancing adversarial robustness. Our results indicate that LGAP outperforms most existing adversarial defense techniques without requiring specialized network training. This underscores the generalizability of models trained on large datasets, highlighting a promising direction for further research.
@inproceedings{singh2024language,
title = {{Language Guided Adversarial Purification}},
author = {Himanshu Singh and A V Subramanyam},
booktitle = {IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
eprint = {2309.10348},
archivePrefix = {arXiv},
doi = {10.1109/icassp48485.2024.10446676},
year = {2024}
}IEEE TCSVT2023
IEEE Transactions on Circuits and Systems for Video Technology
@article{subramanyam2023meta,
title = {{Meta Generative Attack on Person Reidentification}},
author = {A V Subramanyam},
journal = {IEEE Transactions on Circuits and Systems for Video Technology},
eprint = {2301.06286},
archivePrefix = {arXiv},
doi = {10.1109/TCSVT.2023.3237808},
year = {2023}
}IEEE Sensors Letters2023
IEEE Sensors Letters
@article{yasmeen2023circularly,
title = {{Circularly Polarized Fabry-Pèrot Cavity Sensing Antenna Design using Generative Model}},
author = {Kainat Yasmeen and K V Mishra and A V Subramanyam and S S Ram},
journal = {IEEE Sensors Letters},
eprint = {2209.14550},
archivePrefix = {arXiv},
doi = {10.1109/LSENS.2023.3237366},
year = {2023}
}ICVGIP2023
Indian Conference on Computer Vision, Graphics and Image Processing (ICVGIP), pp. 17:1-17:8
Proposes a text-to-image person ReID dataset of 20,000 in-the-wild identities with dense captions, adds Stable Diffusion/BLIP synthetic data, and benchmarks state-of-the-art ReID and vision-language models.
@inproceedings{subramanyam2023dense,
title = {{Dense captioning for Text-Image ReID}},
author = {A V Subramanyam and Vibhu Dubey and Niranjan Sundararajan and Brejesh Lall},
booktitle = {Indian Conference on Computer Vision, Graphics and Image Processing (ICVGIP)},
eprint = {2305.04497},
archivePrefix = {arXiv},
doi = {10.1145/3627631.3627648},
year = {2023}
}ICME2023
IEEE International Conference on Multimedia and Expo (ICME), pp. 2597-2602
Proposes a meta-learning defense for re-identification that trains a stochastic network with anisotropic and isotropic perturbations and a feature covariance alignment loss to resist adversarial attacks.
@inproceedings{verma2023meta,
title = {{Meta Perturbed Re-Id Defense}},
author = {Astha Verma and A V Subramanyam and Mohammad Ali Jauhar and Divij Gera and Rajiv Ratn Shah},
booktitle = {IEEE International Conference on Multimedia and Expo (ICME)},
doi = {10.1109/ICME55011.2023.00442},
year = {2023}
}URSI GASS2023
URSI General Assembly and Scientific Symposium (URSI GASS), Sapporo, Japan
Uses a GAN-based surrogate model, trained on a modest number of antenna designs, to optimize a Fabry-Pérot cavity antenna's reflecting-surface unit cell, validated through simulations and measurements.
@inproceedings{yasmeen2023antenna,
title = {{Antenna Design Optimization using GAN-based Surrogate Model}},
author = {Kainat Yasmeen and Kumar Vijay Mishra and A V Subramanyam and Shobha Sundar Ram},
booktitle = {URSI General Assembly and Scientific Symposium (URSI GASS), Sapporo, Japan},
doi = {10.46620/ursigass.2023.3128.qkzu4883},
year = {2023}
}IEEE TMM2023
IEEE Transactions on Multimedia, vol. 25, pp. 364-377
@article{verma2023unsupervised,
title = {{Unsupervised Domain Adaptation for Person Re-identification via Individual-preserving and Environmental-switching Cyclic Generation}},
author = {A. Verma and A V Subramanyam and Z. Wang and S. Satoh and R. Shah},
journal = {IEEE Transactions on Multimedia},
doi = {10.1109/TMM.2021.3126404},
year = {2023}
}IEEE TIP2022
IEEE Transactions on Image Processing
@article{subramanyam2022sinkhorn,
title = {{Sinkhorn Adversarial Attack and Defense}},
author = {A V Subramanyam},
journal = {IEEE Transactions on Image Processing},
doi = {10.1109/TIP.2022.3180207},
year = {2022}
}ICIP2022
IEEE International Conference on Image Processing (ICIP)
@inproceedings{subramanyam2022barycentric,
title = {{Barycentric Defense}},
author = {A V Subramanyam and Abhigyan Raj},
booktitle = {IEEE International Conference on Image Processing (ICIP)},
doi = {10.1109/ICIP46576.2022.9897732},
year = {2022}
}MIPR2022
IEEE International Conference on Multimedia Information Processing and Retrieval (MIPR)
@inproceedings{verma2022wasserstein,
title = {{Wasserstein Metric Attack on Person Re-Identification}},
author = {Astha Verma and A V Subramanyam and Rajiv Ratn Shah},
booktitle = {IEEE International Conference on Multimedia Information Processing and Retrieval (MIPR)},
doi = {10.1109/MIPR54900.2022.00049},
year = {2022}
}MIPR2022Best Paper Award
IEEE International Conference on Multimedia Information Processing and Retrieval (MIPR)
@inproceedings{joshi2022contextual,
title = {{Contextual Active Learning for Person Re-Identification}},
author = {Vineet Joshi and A V Subramanyam},
booktitle = {IEEE International Conference on Multimedia Information Processing and Retrieval (MIPR)},
doi = {10.1109/MIPR54900.2022.00052},
year = {2022}
}IEEE TCSVT2022
IEEE Transactions on Circuits and Systems for Video Technology, vol. 32, no. 2, pp. 715-729
@article{jain2022channel,
title = {{Channel Graph Regularized Correlation Filters for Visual Object Tracking}},
author = {M. Jain and A. Tyagi and A V Subramanyam and S. Denman and C. Fookes and S. Sridharan},
journal = {IEEE Transactions on Circuits and Systems for Video Technology},
doi = {10.1109/TCSVT.2021.3063144},
year = {2022}
}Wireless Networks2021
Wireless Networks, vol. 27, pp. 4727-4746
@article{chandhok2021novel,
title = {{Novel deep learning framework for wideband spectrum characterization at sub-Nyquist rate}},
author = {S. Chandhok and H. Joshi and A V Subramanyam and S. Darak},
journal = {Wireless Networks},
eprint = {1912.05255},
archivePrefix = {arXiv},
doi = {10.1007/s11276-021-02765-1},
year = {2021}
}WACV2021
IEEE Winter Conference on Applications of Computer Vision (WACV), pp. 2774-2783
@inproceedings{jain2021igsstrcf,
title = {{IGSSTRCF: Importance Guided Sparse Spatio-Temporal Regularized Correlation Filters For Tracking}},
author = {Monika Jain and Arjun Tyagi and A V Subramanyam and Simon Denman and Clinton Fookes and Sridha Sridharan},
booktitle = {IEEE Winter Conference on Applications of Computer Vision (WACV)},
doi = {10.1109/WACV48630.2021.00282},
year = {2021}
}SMC2021
IEEE International Conference on Systems, Man and Cybernetics
@inproceedings{kansal2021disentangling,
title = {{Disentangling Reconstruction Network for Unsupervised Cross-Domain Person Re-Identification}},
author = {Kajal Kansal and Harsh Jain and A V Subramanyam},
booktitle = {IEEE International Conference on Systems, Man and Cybernetics},
doi = {10.1109/SMC52423.2021.9658739},
year = {2021}
}IGARSS2021
IEEE International Geoscience and Remote Sensing Symposium IGARSS, pp. 6650-6653
@inproceedings{shah2021in,
title = {{In-Sat: A Novel Land Cover Classification Dataset for Indian Subcontinent}},
author = {M. Shah and A. V. Subramanyam and G. Arora},
booktitle = {IEEE International Geoscience and Remote Sensing Symposium IGARSS},
doi = {10.1109/IGARSS47720.2021.9553772},
year = {2021}
}ICMEW2021
IEEE International Conference on Multimedia & Expo Workshops (ICMEW), Shenzhen, China, pp. 1-6
@inproceedings{kansal2021hierarchical,
title = {{Hierarchical Attention Image-Text Alignment Network For Person Re-Identification}},
author = {K. Kansal and A. Subramanyam and Z. Wang and S. Satoh},
booktitle = {IEEE International Conference on Multimedia \& Expo Workshops (ICMEW), Shenzhen, China},
doi = {10.1109/ICMEW53276.2021.9455960},
year = {2021}
}CVIU2020
Computer Vision and Image Understanding, vol. 195
@article{jain2020lstm,
title = {{LSTM guided ensemble correlation filter tracking with appearance model pool}},
author = {M. Jain and A V Subramanyam and S. Denman and C. Fookes and S. Sridharan},
journal = {Computer Vision and Image Understanding},
doi = {10.1016/j.cviu.2020.102935},
year = {2020}
}IEEE TCSVT2020
IEEE Transactions on Circuits and Systems for Video Technology, vol. 30, no. 10, pp. 3422-3432, Oct. 2020
@article{kansal2020sdl,
title = {{SDL: Spectrum-Disentangled Representation Learning for Visible-Infrared Person Re-identification}},
author = {K. Kansal and A.V. Subramanyam and Z. Wang and S. Satoh},
journal = {IEEE Transactions on Circuits and Systems for Video Technology},
doi = {10.1109/TCSVT.2019.2963721},
year = {2020}
}IEEE TCSVT2020
IEEE Transactions on Circuits and Systems for Video Technology, vol. 30, no. 9, pp. 3000-3014, Sept. 2020
@article{mehrish2020egocentric,
title = {{Egocentric Analysis of Dash-cam Videos For Vehicle Forensics}},
author = {A. Mehrish and P. Singh and P. Jain and A.V. Subramanyam and M. Kankanhalli},
journal = {IEEE Transactions on Circuits and Systems for Video Technology},
doi = {10.1109/TCSVT.2019.2929561},
year = {2020}
}JVCIR2019
Journal of Visual Communication and Image Representation, vol. 66
@article{sharma2019anti,
title = {{Anti-forensics of Median Filtering and Contrast Enhancement}},
author = {S. Sharma and H. Ravi and A.V. Subramanyam and S. Emmanuel},
journal = {Journal of Visual Communication and Image Representation},
doi = {10.1016/j.jvcir.2019.102682},
year = {2019}
}JEI2019
Journal of Electronic Imaging, vol. 28, no. 2
@article{kansal2019carf,
title = {{CARF-Net: CNN attention and RNN fusion network for video-based person reidentification}},
author = {K. Kansal and A.V. Subramanyam and D. Prasad},
journal = {Journal of Electronic Imaging},
doi = {10.1117/1.JEI.28.2.023036},
year = {2019}
}IEEE Access2019
IEEE Access, vol. 7, pp. 40856-40865
@article{kansal2019hdrnet,
title = {{HDRNet: Person Re-identification using Hybrid Sampling in Deep Reconstruction Network}},
author = {K. Kansal and A.V. Subramanyam and M. Kankanhalli},
journal = {IEEE Access},
doi = {10.1109/ACCESS.2019.2908344},
year = {2019}
}IEEE Access2019
IEEE Access, vol. 7, pp. 27183-27195
@article{mehrish2019joint,
title = {{Joint Spatial and Discrete Cosine Transform Domain based Counter Forensics for Adaptive Contrast Enhancement}},
author = {A. Mehrish and A.V. Subramanyam and S. Emmanuel},
journal = {IEEE Access},
doi = {10.1109/ACCESS.2019.2901345},
year = {2019}
}JEI2019
JEI, vol. 28, no. 2
@article{saini2019airborne,
title = {{Airborne Visual Tracking and Re-Identification System}},
author = {K. Saini and K. Kansal and A.V. Subramanyam},
journal = {JEI},
doi = {10.1117/1.JEI.28.2.023003},
year = {2019}
}JEI2019
Journal of Electronic Imaging, vol. 28, no. 1
@article{sitani2019online,
title = {{Online Single and Multiple Analysis Dictionary Learning based approach for Visual Object Tracking}},
author = {D. Sitani and A.V. Subramanyam and A. Majumdar},
journal = {Journal of Electronic Imaging},
doi = {10.1117/1.JEI.28.1.013004},
year = {2019}
}IEEE SPL2019
IEEE Signal Processing Letters, vol. 26, no. 1, pp. 154-158
@article{subramanyam2019robust,
title = {{Robust Discriminative Sub-space Learning for Person Reidentification}},
author = {A.V. Subramanyam and R. Ahuja and V. Gupta},
journal = {IEEE Signal Processing Letters},
doi = {10.1109/LSP.2018.2882301},
year = {2019}
}ICIP2019
IEEE International Conference on Image Processing (ICIP), pp. 1600-1604 · Poster
@inproceedings{gupta2019attentional,
title = {{Attentional Road Safety Networks}},
author = {Sonu Gupta and Deepak Srivatsav and A V Subramanyam and Ponnurangam Kumaraguru},
booktitle = {IEEE International Conference on Image Processing (ICIP)},
eprint = {1812.04860},
archivePrefix = {arXiv},
doi = {10.1109/ICIP.2019.8803732},
year = {2019}
}BigMM2019
IEEE International Conference on Multimedia Big Data (BigMM), pp. 257-261 · Oral
@inproceedings{kansal2019autoencoder,
title = {{Autoencoder Ensemble for Person Re-Identification}},
author = {Kajal Kansal and A V Subramanyam},
booktitle = {IEEE International Conference on Multimedia Big Data (BigMM)},
doi = {10.1109/BigMM.2019.00-15},
year = {2019}
}BigMM2019
IEEE International Conference on Multimedia Big Data (BigMM), pp. 21-29 · Oral
@inproceedings{singh2019image,
title = {{Image Corpus Representative Summarization}},
author = {Anurag Singh and Lakshay Virmani and A V Subramanyam},
booktitle = {IEEE International Conference on Multimedia Big Data (BigMM)},
doi = {10.1109/BigMM.2019.00-46},
year = {2019}
}COMSNETS2019
International Conference on Communication Systems & Networks (COMSNETS) · Poster
@inproceedings{chandhok2019lstm,
title = {{LSTM Guided Modulation Classification and Experimental Validation for Sub-Nyquist Rate Wideband Spectrum Sensing}},
author = {S. Chandhok and H. Joshi and S. Darak and A.V. Subramanyam},
booktitle = {International Conference on Communication Systems \& Networks (COMSNETS)},
doi = {10.1109/COMSNETS.2019.8711345},
year = {2019}
}Signal Proc.: Image Comm.2018
Signal Processing: Image Communication, vol. 66, pp. 30-41
@article{mehrish2018robust,
title = {{Robust PRNU estimation from probabilistic raw measurements}},
author = {A. Mehrish and A.V. Subramanyam and S. Emmanuel},
journal = {Signal Processing: Image Communication},
doi = {10.1016/j.image.2018.04.013},
year = {2018}
}SMC2018
IEEE International Conference on Systems, Man, and Cybernetics (SMC), pp. 923-928 · Oral
@inproceedings{kansal2018transfer,
title = {{Transfer Learning of Spatio-Temporal Information using 3D-CNN for Person Re-identification}},
author = {K. Kansal and A.V. Subramanyam},
booktitle = {IEEE International Conference on Systems, Man, and Cybernetics (SMC)},
doi = {10.1109/SMC.2018.00164},
year = {2018}
}COMSNETS2018
International Conference on Communication Systems & Networks (COMSNETS), pp. 320-327 · Oral
@inproceedings{banerjee2018use,
title = {{Use of facial landmarks for adaptive compression of videos on mobile devices}},
author = {R. Banerjee and G. Chikara and V. Naik and A.V. Subramanyam and K. Dey},
booktitle = {International Conference on Communication Systems \& Networks (COMSNETS)},
doi = {10.1109/COMSNETS.2018.8328214},
year = {2018}
}ACM MM2017
ACM International Conference on Multimedia (ACM MM), pp. 1434-1442 · Poster
@inproceedings{goel2017visualhashtags,
title = {{\#VisualHashtags: Visual Summarization of Social Media Events Using Mid-Level Visual Elements}},
author = {S. Goel and S. Ahuja and A.V. Subramanyam and P. Kumaraguru},
booktitle = {ACM International Conference on Multimedia (ACM MM)},
doi = {10.1145/3123266.3123407},
year = {2017}
}ICME2017
IEEE International Conference on Multimedia and Expo (ICME), pp. 685-690 · Poster
@inproceedings{mehrish2017multimedia,
title = {{Multimedia Signatures for Vehicle Forensics}},
author = {A. Mehrish and A.V. Subramanyam and M.S. Kankanhalli},
booktitle = {IEEE International Conference on Multimedia and Expo (ICME)},
doi = {10.1109/ICME.2017.8019488},
year = {2017}
}ICIP2017
IEEE International Conference on Image Processing (ICIP), pp. 4327-4331 · Poster
@inproceedings{mullick2017online,
title = {{Online SVM and Backward Model Validation Based Visual Tracking}},
author = {D. Mullick and A.V. Subramanyam and S. Emmanuel},
booktitle = {IEEE International Conference on Image Processing (ICIP)},
doi = {10.1109/ICIP.2017.8297099},
year = {2017}
}IEEE SPL2016
IEEE Signal Processing Letters, vol. 23, no. 5, pp. 693-697
@article{mehrish2016sensor,
title = {{Sensor Pattern Noise Estimation Using Probabilistically Estimated RAW Values}},
author = {A. Mehrish and A.V. Subramanyam and S. Emmanuel},
journal = {IEEE Signal Processing Letters},
doi = {10.1109/LSP.2016.2549059},
year = {2016}
}ACM TOMM2016
ACM Transactions on Multimedia Computing, Communications, and Applications (TOMM), vol. 12, no. 3, pp. 39:1-39:23
@article{ravi2016forensic,
title = {{Forensic Analysis of Linear and Non Linear Image Filtering using Quantization Noise}},
author = {H. Ravi and A.V. Subramanyam and S. Emmanuel},
journal = {ACM Transactions on Multimedia Computing, Communications, and Applications (TOMM)},
doi = {10.1145/2857069},
year = {2016}
}WIFS2016
IEEE International Workshop on Information Forensics and Security (WIFS). Pages: 6 · Oral
@inproceedings{sharma2016anti,
title = {{Anti-Forensic Technique for Median Filtering using L1-L2 TV Model}},
author = {S. Sharma and A.V. Subramanyam and M. Jain and A. Mehrish and S. Emmanuel},
booktitle = {IEEE International Workshop on Information Forensics and Security (WIFS)},
doi = {10.1109/WIFS.2016.7823571},
year = {2016}
}SocInfo2016
International Conference on Social Informatics (SocInfo), pp. 206-223 · Poster
@inproceedings{goel2016pichunt,
title = {{PicHunt: Social Media Image Retrieval for Improved Law Enforcement}},
author = {S. Goel and N. Sachdeva and P. Kumaraguru and A.V. Subramanyam and D. Gupta},
booktitle = {International Conference on Social Informatics (SocInfo)},
eprint = {1608.00905},
archivePrefix = {arXiv},
doi = {10.1007/978-3-319-47880-7_13},
year = {2016}
}MMSP2016
IEEE International Workshop on Multimedia Signal Processing (MMSP). Pages: 5 · Poster
@inproceedings{jain2016face,
title = {{Face Video Based Touchless Blood Pressure and Heart Rate Estimation}},
author = {M. Jain and S. Deb and A.V. Subramanyam},
booktitle = {IEEE International Workshop on Multimedia Signal Processing (MMSP)},
doi = {10.1109/MMSP.2016.7813389},
year = {2016}
}TENCON2016
IEEE Region 10 Conference (TENCON), Singapore, pp. 1361-1365
Proposes a two-stage passive method that uses optical flow consistency to detect whether a video has been inpainted and to temporally localize the forgery, tested on PETS and SULFA.
@inproceedings{saxena2016video,
title = {{Video inpainting detection and localization using inconsistencies in optical flow}},
author = {Shobhita Saxena and A.V. Subramanyam and Hareesh Ravi},
booktitle = {IEEE Region 10 Conference (TENCON), Singapore},
doi = {10.1109/TENCON.2016.7848236},
year = {2016}
}IEEE SPL2015
IEEE Signal Processing Letters, vol. 23, no. 2, pp. 212-216
@article{ravi2015ace,
title = {{ACE - An Effective Anti-forensic Contrast Enhancement Technique}},
author = {H. Ravi and A.V. Subramanyam and S. Emmanuel},
journal = {IEEE Signal Processing Letters},
doi = {10.1109/LSP.2015.2509477},
year = {2015}
}ICIP2015
IEEE International Conference on Image Processing (ICIP), Quebec City, Canada, pp. 1180-1184 · Oral
@inproceedings{ravi2015spatial,
title = {{Spatial domain quantization noise based image filtering detection}},
author = {H. Ravi and A.V. Subramanyam and S. Emmanuel},
booktitle = {IEEE International Conference on Image Processing (ICIP), Quebec City, Canada},
doi = {10.1109/ICIP.2015.7350986},
year = {2015}
}MTAP2014
Multimedia Tools and Applications, vol. 71, no. 3, pp. 1311-1331
@article{subramanyam2014partially,
title = {{Partially Compress-Encrypted Domain Robust JPEG Image Watermarking}},
author = {A.V. Subramanyam and S. Emmanuel},
journal = {Multimedia Tools and Applications},
doi = {10.1007/s11042-012-1272-0},
year = {2014}
}ICIP2014
IEEE International Conference on Image Processing (ICIP), Paris, France, pp. 5353-5356 · Poster
@inproceedings{ravi2014compression,
title = {{Compression Noise Based Video Forgery Detection}},
author = {H. Ravi and A.V. Subramanyam and Gaurav Gupta and B. Avinash Kumar},
booktitle = {IEEE International Conference on Image Processing (ICIP), Paris, France},
doi = {10.1109/ICIP.2014.7026083},
year = {2014}
}SMC2013
IEEE International Conference on Systems, Man and Cybernetics (SMC), Manchester, UK, pp. 1554-1559
@inproceedings{siddula2013cryptanalysis,
title = {{Cryptanalysis of a Digital Watermarking Scheme Based on Support Vector Regression}},
author = {M. Siddula and S. Sanadhya and A.V. Subramanyam},
booktitle = {IEEE International Conference on Systems, Man and Cybernetics (SMC), Manchester, UK},
doi = {10.1109/SMC.2013.268},
year = {2013}
}ICASSP2013
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Vancouver, Canada, pp. 3038-3042 · Poster
@inproceedings{subramanyam2013pixel,
title = {{Pixel Estimation Based Video Forgery Detection}},
author = {A.V. Subramanyam and S. Emmanuel},
booktitle = {IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Vancouver, Canada},
doi = {10.1109/ICASSP.2013.6638216},
year = {2013}
}MMSP2012
IEEE International Workshop on Multimedia Signal Processing (MMSP), Banff, Canada, pp. 89-94 · Poster
@inproceedings{subramanyam2012video,
title = {{Video Forgery Detection using HOG and Video Compression Properties}},
author = {A.V. Subramanyam and S. Emmanuel},
booktitle = {IEEE International Workshop on Multimedia Signal Processing (MMSP), Banff, Canada},
doi = {10.1109/MMSP.2012.6343421},
year = {2012}
}SMC2012
IEEE International Conference on Systems, Man and Cybernetics (SMC), Seoul, South Korea, pp. 2867-2872
@inproceedings{subramanyam2012audio,
title = {{Audio Watermarking in Partially Compressed-Encrypted Domain}},
author = {A.V. Subramanyam and S. Emmanuel},
booktitle = {IEEE International Conference on Systems, Man and Cybernetics (SMC), Seoul, South Korea},
doi = {10.1109/ICSMC.2012.6378184},
year = {2012}
}IEEE TMM2011
IEEE Transactions on Multimedia, vol. 14, no. 3, pp. 703-716
@article{subramanyam2011robust,
title = {{Robust Watermarking of Compressed and Encrypted JPEG2000 Images}},
author = {A.V. Subramanyam and S. Emmanuel and M.S. Kankanhalli},
journal = {IEEE Transactions on Multimedia},
doi = {10.1109/TMM.2011.2181342},
year = {2011}
}LNCS TDHMS2011
Transactions on Data Hiding and Multimedia Security, LNCS, vol. 6730, pp. 37-57
@article{subramanyam2011robustb,
title = {{Robust Watermarking of Compressed JPEG images in Encrypted Domain}},
author = {A.V. Subramanyam and S. Emmanuel},
journal = {Transactions on Data Hiding and Multimedia Security, LNCS},
doi = {10.1007/978-3-642-24556-5_3},
year = {2011}
}SIGMAP2011
International Conference on Signal Processing and Multimedia Applications(SIGMAP), Seville, Spain
@inproceedings{subramanyam2011context,
title = {{Context Based Watermarking of Secure JPEG-LS Images}},
author = {A.V. Subramanyam and S. Emmanuel},
booktitle = {International Conference on Signal Processing and Multimedia Applications(SIGMAP), Seville, Spain},
doi = {10.5220/0003446201610166},
year = {2011}
}ICME2010
IEEE International Conference Multimedia and Expo (ICME), Singapore, pp. 1315-1320 · Poster
@inproceedings{subramanyam2010compressed,
title = {{Compressed Encrypted Domain JPEG2000 Image Watermarking}},
author = {A.V. Subramanyam and S. Emmanuel and M.S. Kankanhalli},
booktitle = {IEEE International Conference Multimedia and Expo (ICME), Singapore},
doi = {10.1109/ICME.2010.5583571},
year = {2010}
}IEEE TIFS2009
IEEE Transactions on Information Forensics and Security, vol. 4, no. 4, pp. 758-767
@article{thomas2009joint,
title = {{Joint Watermarking Scheme for Multiparty Multilevel DRM Architecture}},
author = {T. Thomas and S. Emmanuel and A.V. Subramanyam and M.S. Kankanhalli},
journal = {IEEE Transactions on Information Forensics and Security},
doi = {10.1109/TIFS.2009.2033229},
year = {2009}
}No papers match.