P. J. Narayanan

Sai Sagar Jinka, Astitva Srivastava, Chandradeep Pokhariya et al.

Recent advancements in deep learning have enabled 3D human body reconstruction from a monocular image, which has broad applications in multiple domains. In this paper, we propose SHARP (SHape Aware Reconstruction of People in loose clothing), a novel end-to-end trainable network that accurately recovers the 3D geometry and appearance of humans in loose clothing from a monocular image. SHARP uses a sparse and efficient fusion strategy to combine parametric body prior with a non-parametric 2D representation of clothed humans. The parametric body prior enforces geometrical consistency on the body shape and pose, while the non-parametric representation models loose clothing and handle self-occlusions as well. We also leverage the sparseness of the non-parametric representation for faster training of our network while using losses on 2D maps. Another key contribution is 3DHumans, our new life-like dataset of 3D human body scans with rich geometrical and textural details. We evaluate SHARP on 3DHumans and other publicly available datasets and show superior qualitative and quantitative performance than existing state-of-the-art methods.

Pulkit Gera, Mohammad Reza Karimi Dastjerdi, Charles Renaud et al.

We present PanoHDR-NeRF, a neural representation of the full HDR radiance field of an indoor scene, and a pipeline to capture it casually, without elaborate setups or complex capture protocols. First, a user captures a low dynamic range (LDR) omnidirectional video of the scene by freely waving an off-the-shelf camera around the scene. Then, an LDR2HDR network uplifts the captured LDR frames to HDR, which are used to train a tailored NeRF++ model. The resulting PanoHDR-NeRF can render full HDR images from any location of the scene. Through experiments on a novel test dataset of real scenes with the ground truth HDR radiance captured at locations not seen during training, we show that PanoHDR-NeRF predicts plausible HDR radiance from any scene point. We also show that the predicted radiance can synthesize correct lighting effects, enabling the augmentation of indoor scenes with synthetic objects that are lit correctly. Datasets and code are available at https://lvsn.github.io/PanoHDR-NeRF/.

Rahul Goel, Dhawal Sirikonda, Saurabh Saini et al.

Radiance Fields (RF) are popular to represent casually-captured scenes for new view synthesis and several applications beyond it. Mixed reality on personal spaces needs understanding and manipulating scenes represented as RFs, with semantic segmentation of objects as an important step. Prior segmentation efforts show promise but don't scale to complex objects with diverse appearance. We present the ISRF method to interactively segment objects with fine structure and appearance. Nearest neighbor feature matching using distilled semantic features identifies high-confidence seed regions. Bilateral search in a joint spatio-semantic space grows the region to recover accurate segmentation. We show state-of-the-art results of segmenting objects from RFs and compositing them to another scene, changing appearance, etc., and an interactive segmentation tool that others can use. Project Page: https://rahul-goel.github.io/isrf/

Rahul Goel, Sirikonda Dhawal, Saurabh Saini et al.

Stylized view generation of scenes captured casually using a camera has received much attention recently. The geometry and appearance of the scene are typically captured as neural point sets or neural radiance fields in the previous work. An image stylization method is used to stylize the captured appearance by training its network jointly or iteratively with the structure capture network. The state-of-the-art SNeRF method trains the NeRF and stylization network in an alternating manner. These methods have high training time and require joint optimization. In this work, we present StyleTRF, a compact, quick-to-optimize strategy for stylized view generation using TensoRF. The appearance part is fine-tuned using sparse stylized priors of a few views rendered using the TensoRF representation for a few iterations. Our method thus effectively decouples style-adaption from view capture and is much faster than the previous methods. We show state-of-the-art results on several scenes used for this purpose.

Rahul Goel, Dhawal Sirikonda, Rajvi Shah et al.

Radiance Fields (RFs) have shown great potential to represent scenes from casually captured discrete views. Compositing parts or whole of multiple captured scenes could greatly interest several XR applications. Prior works can generate new views of such scenes by tracing each scene in parallel. This increases the render times and memory requirements with the number of components. In this work, we provide a method to create a single, compact, fused RF representation for a scene composited using multiple RFs. The fused RF has the same render times and memory utilizations as a single RF. Our method distills information from multiple teacher RFs into a single student RF while also facilitating further manipulations like addition and deletion into the fused representation.

Avani Gupta, Saurabh Saini, P J Narayanan

Humans use abstract concepts for understanding instead of hard features. Recent interpretability research has focused on human-centered concept explanations of neural networks. Concept Activation Vectors (CAVs) estimate a model's sensitivity and possible biases to a given concept. In this paper, we extend CAVs from post-hoc analysis to ante-hoc training in order to reduce model bias through fine-tuning using an additional Concept Loss. Concepts were defined on the final layer of the network in the past. We generalize it to intermediate layers using class prototypes. This facilitates class learning in the last convolution layer, which is known to be most informative. We also introduce Concept Distillation to create richer concepts using a pre-trained knowledgeable model as the teacher. Our method can sensitize or desensitize a model towards concepts. We show applications of concept-sensitive training to debias several classification problems. We also use concepts to induce prior knowledge into IID, a reconstruction problem. Concept-sensitive training can improve model interpretability, reduce biases, and induce prior knowledge. Please visit https://avani17101.github.io/Concept-Distilllation/ for code and more details.

Hongjiang Li, Huanyi Shui, Alemayehu Admasu et al.

Image classification with deep neural networks has seen a surge of technological breakthroughs with promising applications in areas such as face recognition, medical imaging, and autonomous driving. In engineering problems, however, such as high-speed imaging of engine fuel injector sprays or body paint sprays, deep neural networks face a fundamental challenge related to the availability of adequate and diverse data. Typically, only thousands or sometimes even hundreds of samples are available for training. In addition, the transition between different spray classes is a continuum and requires a high level of domain expertise to label the images accurately. In this work, we used Mixup as an approach to systematically deal with the data scarcity and ambiguous class boundaries found in industrial spray applications. We show that data augmentation can mitigate the over-fitting problem of large neural networks on small data sets, to a certain level, but cannot fundamentally resolve the issue. We discuss how a convex linear interpolation of different classes naturally aligns with the continuous transition between different classes in our application. Our experiments demonstrate Mixup as a simple yet effective method to train an accurate and robust deep neural network classifier with only a few hundred samples.

Yunfei Long, Daniel Morris, Xiaoming Liu et al.

While radar and video data can be readily fused at the detection level, fusing them at the pixel level is potentially more beneficial. This is also more challenging in part due to the sparsity of radar, but also because automotive radar beams are much wider than a typical pixel combined with a large baseline between camera and radar, which results in poor association between radar pixels and color pixel. A consequence is that depth completion methods designed for LiDAR and video fare poorly for radar and video. Here we propose a radar-to-pixel association stage which learns a mapping from radar returns to pixels. This mapping also serves to densify radar returns. Using this as a first stage, followed by a more traditional depth completion method, we are able to achieve image-guided depth completion with radar and video. We demonstrate performance superior to camera and radar alone on the nuScenes dataset. Our source code is available at https://github.com/longyunf/rc-pda.

Saurabh Saini, P J Narayanan

We present a new additive image factorization technique that treats images to be composed of multiple latent specular components which can be simply estimated recursively by modulating the sparsity during decomposition. Our model-driven {\em RSFNet} estimates these factors by unrolling the optimization into network layers requiring only a few scalars to be learned. The resultant factors are interpretable by design and can be fused for different image enhancement tasks via a network or combined directly by the user in a controllable fashion. Based on RSFNet, we detail a zero-reference Low Light Enhancement (LLE) application trained without paired or unpaired supervision. Our system improves the state-of-the-art performance on standard benchmarks and achieves better generalization on multiple other datasets. We also integrate our factors with other task specific fusion networks for applications like deraining, deblurring and dehazing with negligible overhead thereby highlighting the multi-domain and multi-task generalizability of our proposed RSFNet. The code and data is released for reproducibility on the project homepage.

Avani Gupta, P J Narayanan

The focus of recent research has shifted from merely improving the metrics based performance of Deep Neural Networks (DNNs) to DNNs which are more interpretable to humans. The field of eXplainable Artificial Intelligence (XAI) has observed various techniques, including saliency-based and concept-based approaches. These approaches explain the model's decisions in simple human understandable terms called Concepts. Concepts are known to be the thinking ground of humans}. Explanations in terms of concepts enable detecting spurious correlations, inherent biases, or clever-hans. With the advent of concept-based explanations, a range of concept representation methods and automatic concept discovery algorithms have been introduced. Some recent works also use concepts for model improvement in terms of interpretability and generalization. We provide a systematic review and taxonomy of various concept representations and their discovery algorithms in DNNs, specifically in vision. We also provide details on concept-based model improvement literature marking the first comprehensive survey of these methods.

Pranav Manu, Astitva Srivastava, Amit Raj et al.

Creating photorealistic, animatable, and relightable 3D head avatars traditionally requires expensive Lightstage with multiple calibrated cameras, making it inaccessible for widespread adoption. To bridge this gap, we present a novel, cost-effective approach for creating high-quality relightable head avatars using only a smartphone equipped with polaroid filters. Our approach involves simultaneously capturing cross-polarized and parallel-polarized video streams in a dark room with a single point-light source, separating the skin's diffuse and specular components during dynamic facial performances. We introduce a hybrid representation that embeds 2D Gaussians in the UV space of a parametric head model, facilitating efficient real-time rendering while preserving high-fidelity geometric details. Our learning-based neural analysis-by-synthesis pipeline decouples pose and expression-dependent geometrical offsets from appearance, decomposing the surface into albedo, normal, and specular UV texture maps, along with the environment maps. We collect a unique dataset of various subjects performing diverse facial expressions and head movements.

Venkat Adithya Amula, Sunayana Samavedam, Saurabh Saini et al.

Deep learning models are susceptible to {\em backdoor attacks} involving malicious attackers perturbing a small subset of training data with a {\em trigger} to causes misclassifications. Various triggers have been used, including semantic triggers that are easily realizable without requiring the attacker to manipulate the image. The emergence of generative AI has eased the generation of varied poisoned samples. Robustness across types of triggers is crucial to effective defense. We propose Prototype Guided Backdoor Defense (PGBD), a robust post-hoc defense that scales across different trigger types, including previously unsolved semantic triggers. PGBD exploits displacements in the geometric spaces of activations to penalize movements toward the trigger. This is done using a novel sanitization loss of a post-hoc fine-tuning step. The geometric approach scales easily to all types of attacks. PGBD achieves better performance across all settings. We also present the first defense against a new semantic attack on celebrity face images. Project page: \hyperlink{https://venkatadithya9.github.io/pgbd.github.io/}{this https URL}.

Vinayak Gupta, Rahul Goel, Sirikonda Dhawal et al.

Traditional Radiance Field (RF) representations capture details of a specific scene and must be trained afresh on each scene. Semantic feature fields have been added to RFs to facilitate several segmentation tasks. Generalised RF representations learn the principles of view interpolation. A generalised RF can render new views of an unknown and untrained scene, given a few views. We present a way to distil feature fields into the generalised GNT representation. Our GSN representation generates new views of unseen scenes on the fly along with consistent, per-pixel semantic features. This enables multi-view segmentation of arbitrary new scenes. We show different semantic features being distilled into generalised RFs. Our multi-view segmentation results are on par with methods that use traditional RFs. GSN closes the gap between standard and generalisable RF methods significantly. Project Page: https://vinayak-vg.github.io/GSN/

Pulkit Gera, Aakash KT, Dhawal Sirikonda et al.

We present a neural rendering framework for simultaneous view synthesis and appearance editing of a scene from multi-view images captured under known environment illumination. Existing approaches either achieve view synthesis alone or view synthesis along with relighting, without direct control over the scene's appearance. Our approach explicitly disentangles the appearance and learns a lighting representation that is independent of it. Specifically, we independently estimate the BRDF and use it to learn a lighting-only representation of the scene. Such disentanglement allows our approach to generalize to arbitrary changes in appearance while performing view synthesis. We show results of editing the appearance of a real scene, demonstrating that our approach produces plausible appearance editing. The performance of our view synthesis approach is demonstrated to be at par with state-of-the-art approaches on both real and synthetic data.

Yunfei Long, Daniel Morris, Xiaoming Liu et al.

A distinctive feature of Doppler radar is the measurement of velocity in the radial direction for radar points. However, the missing tangential velocity component hampers object velocity estimation as well as temporal integration of radar sweeps in dynamic scenes. Recognizing that fusing camera with radar provides complementary information to radar, in this paper we present a closed-form solution for the point-wise, full-velocity estimate of Doppler returns using the corresponding optical flow from camera images. Additionally, we address the association problem between radar returns and camera images with a neural network that is trained to estimate radar-camera correspondences. Experimental results on the nuScenes dataset verify the validity of the method and show significant improvements over the state-of-the-art in velocity estimation and accumulation of radar points.

Sai Sagar Jinka, Rohan Chacko, Astitva Srivastava et al.

3D human body reconstruction from monocular images is an interesting and ill-posed problem in computer vision with wider applications in multiple domains. In this paper, we propose SHARP, a novel end-to-end trainable network that accurately recovers the detailed geometry and appearance of 3D people in loose clothing from a monocular image. We propose a sparse and efficient fusion of a parametric body prior with a non-parametric peeled depth map representation of clothed models. The parametric body prior constraints our model in two ways: first, the network retains geometrically consistent body parts that are not occluded by clothing, and second, it provides a body shape context that improves prediction of the peeled depth maps. This enables SHARP to recover fine-grained 3D geometrical details with just L1 losses on the 2D maps, given an input image. We evaluate SHARP on publicly available Cloth3D and THuman datasets and report superior performance to state-of-the-art approaches.

Parikshit Sakurikar, P. J. Narayanan

An image captured with a wide-aperture camera exhibits a finite depth-of-field, with focused and defocused pixels. A compact and robust representation of focus and defocus helps analyze and manipulate such images. In this work, we study the fine characteristics of images with a shallow depth-of-field in the context of focal stacks. We present a composite measure for focus that is a combination of existing measures. We identify in-focus pixels, dual-focus pixels, pixels that exhibit bokeh and spatially-varying blur kernels between focal slices. We use these to build a novel representation that facilitates easy manipulation of focal stacks. We present a comprehensive algorithm for post-capture refocusing in a geometrically correct manner. Our approach can refocus the scene at high fidelity while preserving fine aspects of focus and defocus blur.

Sai Sagar Jinka, Rohan Chacko, Avinash Sharma et al.

We introduce PeeledHuman - a novel shape representation of the human body that is robust to self-occlusions. PeeledHuman encodes the human body as a set of Peeled Depth and RGB maps in 2D, obtained by performing ray-tracing on the 3D body model and extending each ray beyond its first intersection. This formulation allows us to handle self-occlusions efficiently compared to other representations. Given a monocular RGB image, we learn these Peeled maps in an end-to-end generative adversarial fashion using our novel framework - PeelGAN. We train PeelGAN using a 3D Chamfer loss and other 2D losses to generate multiple depth values per-pixel and a corresponding RGB field per-vertex in a dual-branch setup. In our simple non-parametric solution, the generated Peeled Depth maps are back-projected to 3D space to obtain a complete textured 3D shape. The corresponding RGB maps provide vertex-level texture details. We compare our method with current parametric and non-parametric methods in 3D reconstruction and find that we achieve state-of-the-art-results. We demonstrate the effectiveness of our representation on publicly available BUFF and MonoPerfCap datasets as well as loose clothing data collected by our calibrated multi-Kinect setup.

Nikita Jaipuria, Shubh Gupta, Praveen Narayanan et al.

Generative Adversarial Networks (GANs) are now widely used for photo-realistic image synthesis. In applications where a simulated image needs to be translated into a realistic image (sim-to-real), GANs trained on unpaired data from the two domains are susceptible to failure in semantic content retention as the image is translated from one domain to the other. This failure mode is more pronounced in cases where the real data lacks content diversity, resulting in a content \emph{mismatch} between the two domains - a situation often encountered in real-world deployment. In this paper, we investigate the role of the discriminator's receptive field in GANs for unsupervised image-to-image translation with mismatched data, and study its effect on semantic content retention. Experiments with the discriminator architecture of a state-of-the-art coupled Variational Auto-Encoder (VAE) - GAN model on diverse, mismatched datasets show that the discriminator receptive field is directly correlated with semantic content discrepancy of the generated image.

Praveen Narayanan, Punarjay Chakravarty, Francois Charette et al.

We present a voice conversion solution using recurrent sequence to sequence modeling for DNNs. Our solution takes advantage of recent advances in attention based modeling in the fields of Neural Machine Translation (NMT), Text-to-Speech (TTS) and Automatic Speech Recognition (ASR). The problem consists of converting between voices in a parallel setting when {\it $<$source,target$>$} audio pairs are available. Our seq2seq architecture makes use of a hierarchical encoder to summarize input audio frames. On the decoder side, we use an attention based architecture used in recent TTS works. Since there is a dearth of large multispeaker voice conversion databases needed for training DNNs, we resort to training the network with a large single speaker dataset as an autoencoder. This is then adapted for the smaller multispeaker voice conversion datasets available for voice conversion. In contrast with other voice conversion works that use $F_0$, duration and linguistic features, our system uses mel spectrograms as the audio representation. Output mel frames are converted back to audio using a wavenet vocoder.

Saurabh Saini, P. J. Narayanan

Intrinsic Image Decomposition (IID) is a challenging and interesting computer vision problem with various applications in several fields. We present novel semantic priors and an integrated approach for single image IID that involves analyzing image at three hierarchical context levels. Local context priors capture scene properties at each pixel within a small neighbourhood. Mid-level context priors encode object level semantics. Global context priors establish correspondences at the scene level. Our semantic priors are designed on both fixed and flexible regions, using selective search method and Convolutional Neural Network features. Our IID method is an iterative multistage optimization scheme and consists of two complementary formulations: $L_2$ smoothing for shading and $L_1$ sparsity for reflectance. Experiments and analysis of our method indicate the utility of our semantic priors and structured hierarchical analysis in an IID framework. We compare our method with other contemporary IID solutions and show results with lesser artifacts. Finally, we highlight that proper choice and encoding of prior knowledge can produce competitive results even when compared to end-to-end deep learning IID methods, signifying the importance of such priors. We believe that the insights and techniques presented in this paper would be useful in the future IID research.

Punarjay Chakravarty, Praveen Narayanan, Tom Roussel

We present a Deep Learning based system for the twin tasks of localization and obstacle avoidance essential to any mobile robot. Our system learns from conventional geometric SLAM, and outputs, using a single camera, the topological pose of the camera in an environment, and the depth map of obstacles around it. We use a CNN to localize in a topological map, and a conditional VAE to output depth for a camera image, conditional on this topological location estimation. We demonstrate the effectiveness of our monocular localization and depth estimation system on simulated and real datasets.

Aryaman Gupta, Kalpit Thakkar, Vineet Gandhi et al.

Monocular head pose estimation requires learning a model that computes the intrinsic Euler angles for pose (yaw, pitch, roll) from an input image of human face. Annotating ground truth head pose angles for images in the wild is difficult and requires ad-hoc fitting procedures (which provides only coarse and approximate annotations). This highlights the need for approaches which can train on data captured in controlled environment and generalize on the images in the wild (with varying appearance and illumination of the face). Most present day deep learning approaches which learn a regression function directly on the input images fail to do so. To this end, we propose to use a higher level representation to regress the head pose while using deep learning architectures. More specifically, we use the uncertainty maps in the form of 2D soft localization heatmap images over five facial keypoints, namely left ear, right ear, left eye, right eye and nose, and pass them through an convolutional neural network to regress the head-pose. We show head pose estimation results on two challenging benchmarks BIWI and AFLW and our approach surpasses the state of the art on both the datasets.

Kalpit Thakkar, P J Narayanan

Human actions comprise of joint motion of articulated body parts or `gestures'. Human skeleton is intuitively represented as a sparse graph with joints as nodes and natural connections between them as edges. Graph convolutional networks have been used to recognize actions from skeletal videos. We introduce a part-based graph convolutional network (PB-GCN) for this task, inspired by Deformable Part-based Models (DPMs). We divide the skeleton graph into four subgraphs with joints shared across them and learn a recognition model using a part-based graph convolutional network. We show that such a model improves performance of recognition, compared to a model using entire skeleton graph. Instead of using 3D joint coordinates as node features, we show that using relative coordinates and temporal displacements boosts performance. Our model achieves state-of-the-art performance on two challenging benchmark datasets NTURGB+D and HDM05, for skeletal action recognition.

Rajvi Shah, Visesh Chari, P J Narayanan

View-graph is an essential input to large-scale structure from motion (SfM) pipelines. Accuracy and efficiency of large-scale SfM is crucially dependent on the input view-graph. Inconsistent or inaccurate edges can lead to inferior or wrong reconstruction. Most SfM methods remove `undesirable' images and pairs using several, fixed heuristic criteria, and propose tailor-made solutions to achieve specific reconstruction objectives such as efficiency, accuracy, or disambiguation. In contrast to these disparate solutions, we propose a single optimization framework that can be used to achieve these different reconstruction objectives with task-specific cost modeling. We also construct a very efficient network-flow based formulation for its approximate solution. The abstraction brought on by this selection mechanism separates the challenges specific to datasets and reconstruction objectives from the standard SfM pipeline and improves its generalization. This paper demonstrates the application of the proposed view-graph framework with standard SfM pipeline for two particular use-cases, (i) accurate and ghost-free reconstructions of highly ambiguous datasets using costs based on disambiguation priors, and (ii) accurate and efficient reconstruction of large-scale Internet datasets using costs based on commonly used priors.

Aditya Singh, Saurabh Saini, Rajvi Shah et al.

User-given tags or labels are valuable resources for semantic understanding of visual media such as images and videos. Recently, a new type of labeling mechanism known as hash-tags have become increasingly popular on social media sites. In this paper, we study the problem of generating relevant and useful hash-tags for short video clips. Traditional data-driven approaches for tag enrichment and recommendation use direct visual similarity for label transfer and propagation. We attempt to learn a direct low-cost mapping from video to hash-tags using a two step training process. We first employ a natural language processing (NLP) technique, skip-gram models with neural network training to learn a low-dimensional vector representation of hash-tags (Tag2Vec) using a corpus of 10 million hash-tags. We then train an embedding function to map video features to the low-dimensional Tag2vec space. We learn this embedding for 29 categories of short video clips with hash-tags. A query video without any tag-information can then be directly mapped to the vector space of tags using the learned embedding and relevant tags can be found by performing a simple nearest-neighbor retrieval in the Tag2Vec space. We validate the relevance of the tags suggested by our system qualitatively and quantitatively with a user study.

Aditya Singh, Saurabh Saini, Rajvi Shah et al.

Short internet video clips like vines present a significantly wild distribution compared to traditional video datasets. In this paper, we focus on the problem of unsupervised action classification in wild vines using traditional labeled datasets. To this end, we use a data augmentation based simple domain adaptation strategy. We utilise semantic word2vec space as a common subspace to embed video features from both, labeled source domain and unlablled target domain. Our method incrementally augments the labeled source with target samples and iteratively modifies the embedding function to bring the source and target distributions together. Additionally, we utilise a multi-modal representation that incorporates noisy semantic information available in form of hash-tags. We show the effectiveness of this simple adaptation technique on a test set of vines and achieve notable improvements in performance.

Rajvi Shah, Aditya Deshpande, P J Narayanan

Several methods have been proposed for large-scale 3D reconstruction from large, unorganized image collections. A large reconstruction problem is typically divided into multiple components which are reconstructed independently using structure from motion (SFM) and later merged together. Incremental SFM methods are most popular for the basic structure recovery of a single component. They are robust and effective but are strictly sequential in nature. We present a multistage approach for SFM reconstruction of a single component that breaks the sequential nature of the incremental SFM methods. Our approach begins with quickly building a coarse 3D model using only a fraction of features from given images. The coarse model is then enriched by localizing remaining images and matching and triangulating remaining features in subsequent stages. These stages are made efficient and highly parallel by leveraging the geometry of the coarse model. Our method produces similar quality models as compared to incremental SFM methods while being notably fast and parallel.