1 code implementation • 25 Oct 2023 • Prajwal Singh, Dwip Dalal, Gautam Vashishtha, Krishna Miyapuram, Shanmuganathan Raman
Reconstruction of visual images from brain Electroencephalography (EEG) signals has garnered a lot of interest due to its applications in brain-computer interfacing.
no code implementations • 6 Jul 2023 • Dwip Dalal, Gautam Vashishtha, Prajwal Singh, Shanmuganathan Raman
Digital imaging aims to replicate realistic scenes, but Low Dynamic Range (LDR) cameras cannot represent the wide dynamic range of real scenes, resulting in under-/overexposed images.
no code implementations • 23 Jun 2023 • Aalok Gangopadhyay, Abhinav Narayan Harish, Prajwal Singh, Shanmuganathan Raman
We have proposed a self-supervised deep learning framework for solving the mesh blending problem in scenarios where the meshes are not in correspondence.
no code implementations • 20 Nov 2022 • Aadesh Desai, Saagar Parikh, Seema Kumari, Shanmuganathan Raman
Point cloud segmentation and classification are some of the primary tasks in 3D computer vision with applications ranging from augmented reality to robotics.
no code implementations • 1 Nov 2022 • Ashish Tiwari, Sresth Tosniwal, Shanmuganathan Raman
Graph convolution networks (GCNs) have been enormously successful in learning representations over several graph-based machine learning tasks.
no code implementations • 10 Oct 2022 • Zeeshan Khan, Parth Shettiwar, Mukul Khanna, Shanmuganathan Raman
Previous works in high dynamic range (HDR) video reconstruction uses sequence of alternating exposure LDR frames as input, and align the neighbouring frames using optical flow based networks.
no code implementations • 5 Jul 2022 • Ashish Tiwari, Shanmuganathan Raman
Despite the success of existing traditional and deep learning-based methods, it is still challenging due to: (i) the requirement of three or more differently illuminated images, (ii) the inability to model unknown general reflectance, and (iii) the requirement of accurate 3D ground truth surface normals and known lighting information for training.
no code implementations • 30 Nov 2021 • Jatin Kumar, Indra Deep Mastan, Shanmuganathan Raman
With the help of MobileNet based architecture that consists of depthwise separable convolution, we reduce the model size and inference time, without losing the quality of the images.
no code implementations • 22 Oct 2021 • Mrinal Anand, Nidhin Harilal, Chandan Kumar, Shanmuganathan Raman
We first extract clean LDR frames from noisy LDR video with alternating exposures with a denoising network trained in a self-supervised setting.
1 code implementation • 7 Oct 2021 • Prajwal Singh, Kaustubh Sadekar, Shanmuganathan Raman
Point cloud is one of the widely used techniques for representing and storing 3D geometric data.
no code implementations • 4 Oct 2021 • Prarabdh Raipurkar, Rohil Pal, Shanmuganathan Raman
Specifically, saturation in overexposed regions makes the task of reconstructing a High Dynamic Range (HDR) image from single LDR image challenging.
no code implementations • 30 Jul 2021 • Kaustubh Sadekar, Ashish Tiwari, Shanmuganathan Raman
In this work, we revisit shadow art using differentiable rendering based optimization frameworks to obtain the 3D sculpture from a set of shadow (binary) images and their corresponding projection information.
1 code implementation • 27 Jul 2021 • Abhinav Narayan Harish, Rajendra Nagar, Shanmuganathan Raman
Autonomous assembly of objects is an essential task in robotics and 3D computer vision.
no code implementations • 17 Jun 2021 • Sudhakar Kumawat, Gagan Kanojia, Shanmuganathan Raman
This paper studies the operation of channel shuffle as a regularization technique in deep convolutional networks.
no code implementations • 27 Jan 2021 • Kaustubh Sadekar, Ashish Tiwari, Prajwal Singh, Shanmuganathan Raman
This work proposes LS-HDIB - a large-scale handwritten document image binarization dataset containing over a million document images that span numerous real-world scenarios.
1 code implementation • 15 Jan 2021 • Aalok Gangopadhyay, Prajwal Singh, Shanmuganathan Raman
To minimize this intervention, we propose APEX-Net, a deep learning based framework with novel loss functions for solving the plot extraction problem.
no code implementations • 11 Dec 2020 • Indra Deep Mastan, Shanmuganathan Raman
Recent methods for image enhancement consider the problem by performing style transfer and image restoration.
no code implementations • 11 Dec 2020 • Indra Deep Mastan, Shanmuganathan Raman
DeepObjStyle preserves the semantics of the objects and achieves better style transfer in the challenging scenario when the style and the content images have a mismatch of image features.
no code implementations • 7 Nov 2020 • Indra Deep Mastan, Shanmuganathan Raman
In this work, we proposed a new training data-independent framework, called Deep Contextual Features Learning (DeepCFL), to perform image synthesis and image restoration based on the semantics of the input image.
no code implementations • 7 Nov 2020 • Harshil Jain, Rohit Patil, Indra Deep Mastan, Shanmuganathan Raman
SinGAN is a generative model that is unconditional and could be learned from a single natural image.
no code implementations • 22 Oct 2020 • Gagan Kanojia, Shanmuganathan Raman
Let the temporal order in which these images are captured be unknown.
no code implementations • 22 Jul 2020 • Sudhakar Kumawat, Manisha Verma, Yuta Nakashima, Shanmuganathan Raman
To address these issues, we propose spatio-temporal short term Fourier transform (STFT) blocks, a new class of convolutional blocks that can serve as an alternative to the 3D convolutional layer and its variants in 3D CNNs.
1 code implementation • 22 Apr 2020 • Manisha Verma, Sudhakar Kumawat, Yuta Nakashima, Shanmuganathan Raman
To handle more variety in human poses, we propose the concept of fine-grained hierarchical pose classification, in which we formulate the pose estimation as a classification task, and propose a dataset, Yoga-82, for large-scale yoga pose recognition with 82 classes.
no code implementations • 8 Feb 2020 • Chandra Sekhar Ravuri, Rajesh Sureddi, Sathya Veera Reddy Dendi, Shanmuganathan Raman, Sumohana S. Channappayya
The novelty of this work is its ability to visualize various distortions as quality maps (distortion maps), especially in the no-reference setting, and to use these maps as features to estimate the quality score of tone mapped images.
no code implementations • 27 Jan 2020 • Sudhakar Kumawat, Shanmuganathan Raman
In this paper, we propose a new convolutional layer called Depthwise-STFT Separable layer that can serve as an alternative to the standard depthwise separable convolutional layer.
1 code implementation • 24 Dec 2019 • Zeeshan Khan, Mukul Khanna, Shanmuganathan Raman
High dynamic range (HDR) image generation from a single exposure low dynamic range (LDR) image has been made possible due to the recent advances in Deep Learning.
no code implementations • 11 Dec 2019 • Gagan Kanojia, Shanmuganathan Raman
During the scan, when a pixel is classified as dynamic, the proposed algorithm replaces that pixel value with the corresponding pixel value of the static region which is being occluded by that dynamic region.
no code implementations • 9 Dec 2019 • Indra Deep Mastan, Shanmuganathan Raman
Recently, there is a vast interest in developing methods which are independent of the training samples such as deep image prior, zero-shot learning, and internal learning.
no code implementations • 18 Nov 2019 • Shubham Kumar Singh, Krishna P. Miyapuram, Shanmuganathan Raman
This model factorizes the visual appearance of a person into latent discriminative factors at multiple semantic levels.
no code implementations • 7 Sep 2019 • Gagan Kanojia, Sudhakar Kumawat, Shanmuganathan Raman
Traditional 3D convolutions are computationally expensive, memory intensive, and due to large number of parameters, they often tend to overfit.
no code implementations • 1 May 2019 • Indra Deep Mastan, Shanmuganathan Raman
In this paper, we propose a framework based on the multi-level extensions of the encoder-decoder network, to investigate interesting aspects of the relationship between image restoration and network construction independent of learning.
no code implementations • 30 Apr 2019 • Gagan Kanojia, Sudhakar Kumawat, Shanmuganathan Raman
The proposed model outperforms the classification accuracy of the state-of-the-art models in both 2D and 3D frameworks by 11. 54% and 4. 24%, respectively.
no code implementations • 16 Apr 2019 • Sudhakar Kumawat, Manisha Verma, Shanmuganathan Raman
Recognizing facial expressions is one of the central problems in computer vision.
Facial Expression Recognition Facial Expression Recognition (FER)
no code implementations • CVPR 2019 • Sudhakar Kumawat, Shanmuganathan Raman
The ReLPV block extracts the phase in a 3D local neighborhood (e. g., 3x3x3) of each position of the input map to obtain the feature maps.
no code implementations • ECCV 2018 • Rajendra Nagar, Shanmuganathan Raman
In this work, we detect the intrinsic reflective symmetry in triangle meshes where we have to find the intrinsically symmetric point for each point of the shape.
no code implementations • 23 May 2018 • Rajendra Nagar, Shanmuganathan Raman
We partition the image into superpixels while preserving this reflection symmetry through an iterative algorithm.
1 code implementation • 11 Sep 2017 • Sainandan Ramakrishnan, Shubham Pachori. Aalok Gangopadhyay, Shanmuganathan Raman
Removing blur caused by camera shake in images has always been a challenging problem in computer vision literature due to its ill-posed nature.
no code implementations • 26 Aug 2017 • Viraj Mavani, Shanmuganathan Raman, Krishna P. Miyapuram
We have developed a convolutional neural network for the purpose of recognizing facial expressions in human beings.
Facial Expression Recognition Facial Expression Recognition (FER) +1
no code implementations • 2 Jul 2017 • Vikas Gupta, Shanmuganathan Raman
We use learning based matting method to generate the matte from the automatically generated trimap.
no code implementations • 29 Jun 2017 • Aditya Vora, Shanmuganathan Raman
Our algorithm segments various object instances appearing in the video and then perform clustering in order to group visually similar segments into one cluster.
no code implementations • 29 Jun 2017 • Aditya Vora, Shanmuganathan Raman
This paper addresses the problem of unsupervised object localization in an image.
no code implementations • 27 Jun 2017 • Rajendra Nagar, Shanmuganathan Raman
We formulate an optimization framework in which the problem of establishing the correspondences amounts to solving a linear assignment problem and the problem of determining the reflection symmetry transformation amounts to solving an optimization problem on a smooth Riemannian product manifold.
no code implementations • 7 Feb 2017 • Shubham Pachori, Ameya Deshpande, Shanmuganathan Raman
Therefore, we propose an algorithm to learn a hash function from training images belonging to `seen' classes which can efficiently encode images of `unseen' classes to binary codes.
no code implementations • 9 Oct 2016 • Shubham Pachori, Shanmuganathan Raman
In this work, we attempt to generate the hash codes for images belonging to unseen classes, information of which is available only through the textual corpus.
no code implementations • 16 Apr 2016 • Sri Raghu Malireddi, Shanmuganathan Raman
Automatic segmentation of objects from a single image is a challenging problem which generally requires training on large number of images.
no code implementations • 7 Apr 2016 • Akshay Gadi Patil, Shanmuganathan Raman
Non-photorealistic rendering techniques work on image features and often manipulate a set of characteristics such as edges and texture to achieve a desired depiction of the scene.
no code implementations • 2 Oct 2015 • Ishan Jindal, Shanmuganathan Raman
The analysis shows the advantages and limitations of the proposed approach for tracking an object in unstructured crowd scenes.
no code implementations • 17 Feb 2015 • Aalok Gangopadhyay, Shivam Mani Tripathi, Ishan Jindal, Shanmuganathan Raman
The task of classifying videos of natural dynamic scenes into appropriate classes has gained lot of attention in recent years.
no code implementations • 20 May 2013 • Govind Salvi, Puneet Sharma, Shanmuganathan Raman
In this paper, we address the problem of displaying the high contrast low dynamic range (LDR) image of a HDR scene in a display device which has different spatial resolution compared to that of the capturing digital camera.