no code implementations • 9 Jul 2023 • Ayush Singh, Yash Bhambhu, Himanshu Buckchash, Deepak K. Gupta, Dilip K. Prasad
In this paper, we present Latent Graph Attention (LGA) a computationally inexpensive (linear to the number of nodes) and stable, modular framework for incorporating the global context in the existing architectures, especially empowering small-scale architectures to give performance closer to large size architectures, thus making the light-weight architectures more useful for edge devices with lower compute power and lower energy needs.
1 code implementation • 3 Mar 2023 • Animesh Gupta, Irtiza Hasan, Dilip K. Prasad, Deepak K. Gupta
We further show that when no pretraining is done or when the pretrained transformer models are used with non-natural images (e. g. medical data), CNNs tend to generalize better than transformers at even very small coreset sizes.
Ranked #3 on Image Classification on Tiny ImageNet Classification
no code implementations • 31 Jan 2023 • Deepak K. Gupta, Gowreesh Mago, Arnav Chavan, Dilip K. Prasad
Traditional CNN models are trained and tested on relatively low resolution images (<300 px), and cannot be directly operated on large-scale images due to compute and memory constraints.
1 code implementation • 16 Jan 2023 • Fatima Albreiki, Nidhal Belayouni, Deepak K. Gupta
The correctness of the approximate model depends on the extent of sampling conducted in the parameter space, and through numerical experiments, we demonstrate that caution needs to be taken when constructing this landscape with neural networks.
1 code implementation • 24 Nov 2022 • Saksham Aggarwal, Taneesh Gupta, Pawan Kumar Sahu, Arnav Chavan, Rishabh Tiwari, Dilip K. Prasad, Deepak K. Gupta
A comparison between SOTA trackers using CNNs, transformers as well as the combination of the two is presented to study their stability at various compression ratios.
no code implementations • 12 Nov 2022 • Udbhav Bamba, Neeraj Anand, Saksham Aggarwal, Dilip K. Prasad, Deepak K. Gupta
To address this issue, partial binarization techniques have been developed, but a systematic approach to mixing binary and full-precision parameters in a single network is still lacking.
no code implementations • 25 Jun 2022 • Deepak K. Gupta, Udbhav Bamba, Abhishek Thakur, Akash Gupta, Suraj Sharan, Ertugrul Demir, Dilip K. Prasad
Based on the outlined issues, we introduce a novel research problem of training CNN models for very large images, and present 'UltraMNIST dataset', a simple yet representative benchmark dataset for this task.
1 code implementation • CVPR 2022 • Arnav Chavan, Rishabh Tiwari, Udbhav Bamba, Deepak K. Gupta
MetaDOCK compresses the meta-model as well as the task-specific inner models, thus providing significant reduction in model size for each task, and through constraining the number of active kernels for every task, it implicitly mitigates the issue of meta-overfitting.
1 code implementation • 28 Nov 2021 • Naman Khetan, Tushar Arora, Samee Ur Rehman, Deepak K. Gupta
Several approaches exist that make CNNs equivariant under other transformation groups by design.
no code implementations • 1 Nov 2021 • Bhavesh Tangirala, Ishan Bhandari, Daniel Laszlo, Deepak K. Gupta, Rajat M. Thomas, Devanshu Arya
Given these challenges, we develop an end-to-end behaviour monitoring system for group-housed pigs to perform simultaneous instance level segmentation, tracking, action recognition and re-identification (STAR) tasks.
no code implementations • 22 Sep 2021 • Devanshu Arya, Deepak K. Gupta, Stevan Rudinac, Marcel Worring
Most hypergraph learning approaches convert the hypergraph structure to that of a graph and then deploy existing geometric deep learning methods.
1 code implementation • ICLR 2021 • Rishabh Tiwari, Udbhav Bamba, Arnav Chavan, Deepak K. Gupta
Structured pruning methods are among the effective strategies for extracting small resource-efficient convolutional neural networks from their dense counterparts with minimal loss in accuracy.
2 code implementations • CVPR 2021 • Deepak K. Gupta, Devanshu Arya, Efstratios Gavves
We further show that this change in orientation can be used to impose an additional motion constraint in Siamese tracking through imposing restriction on the change in orientation between two consecutive frames.
no code implementations • 9 Oct 2020 • Devanshu Arya, Deepak K. Gupta, Stevan Rudinac, Marcel Worring
To model such complex relations, hypergraphs have proven to be a natural representation.
no code implementations • 10 Sep 2020 • Thijs P. Kuipers, Devanshu Arya, Deepak K. Gupta
A tracker is assessed to be good or not based on its performance on the recent tracking datasets, e. g., VOT2019, and LaSOT.
no code implementations • 30 Jun 2020 • Deepak K. Gupta, Efstratios Gavves, Arnold W. M. Smeulders
Specifically, we present structured dropout to mimick the change in latent codes under occlusion.
3 code implementations • ICCV 2021 • Elias Kassapis, Georgi Dikov, Deepak K. Gupta, Cedric Nugteren
To this end, we propose a novel two-stage, cascaded approach for calibrated adversarial refinement: (i) a standard segmentation network is trained with categorical cross entropy to predict a pixelwise probability distribution over semantic classes and (ii) an adversarially trained stochastic network is used to model the inter-pixel correlations to refine the output of the first network into coherent samples.
1 code implementation • 22 Jun 2020 • Bryan G. Cardenas, Devanshu Arya, Deepak K. Gupta
In particular, layout-to-image generation models have gained significant attention due to their capability to generate realistic complex images containing distinct objects.
Layout-to-Image Generation Vocal Bursts Intensity Prediction
no code implementations • MIDL 2019 • Devanshu Arya, Richard Olij, Deepak K. Gupta, Ahmed El Gazzar, Guido van Wingen, Marcel Worring, Rajat Mani Thomas
We alleviate the use of such non-imaging metadata and propose a fully imaging-based approach where information from structural and functional Magnetic Resonance Imaging (MRI) data are fused to construct the edges and nodes of the graph.
no code implementations • MIDL 2019 • Andreas Panteli, Deepak K. Gupta, Nathan de Bruin, Efstratios Gavves
Tracking and segmentation of biological cells in video sequences is a challenging problem, especially due to the similarity of the cells and high levels of inherent noise.
no code implementations • 19 Oct 2019 • Deepak K. Gupta, Nathan de Bruijn, Andreas Panteli, Efstratios Gavves
U-Net and its variants have been demonstrated to work sufficiently well in biological cell tracking and segmentation.
no code implementations • 5 Aug 2019 • Efstratios Gavves, Ran Tao, Deepak K. Gupta, Arnold W. M. Smeulders
Updating the tracker model with adverse bounding box predictions adds an unavoidable bias term to the learning.
no code implementations • 22 Jan 2019 • Deepak K. Gupta, Rohit K. Shrivastava, Suhas Phadke, Jeroen Goudswaard
The ICS approach consists of a few hyper-parameters that have been chosen based on statistical study performed over a set of test images.