Abstract

Multi-object multi-part scene parsing is a challenging task which requires detecting multiple object classes in a scene and segmenting the semantic parts within each object. In this paper, we propose FLOAT, a factorized label space framework for scalable multi-object multi-part parsing. Our framework involves independent dense prediction of object category and part attributes which increases scalability and reduces task complexity compared to the monolithic label space counterpart. In addition, we propose an inference-time ‘zoom’ refinement technique which significantly improves segmentation quality, especially for smaller objects/parts. Compared to state of the art, FLOAT obtains an absolute improvement of 2.0% for mean IOU (mIOU) and 4.8% for segmentation quality IOU (sqIOU) on the Pascal-Part-58 dataset. For the larger Pascal-Part-108 dataset, the improvements are 2.1% for mIOU and 3.9% for

Abstract

Linearly Transformed Cosines (LTCs) are a family of distributions that are used for real-time area-light shading thanks to their analytic integration properties. Modern game engines use an LTC approximation of the ubiquitous GGX model, but currently this approximation only exists for isotropic GGX and thus anisotropic GGX is not supported. While the higher dimensionality presents a challenge in itself, we show that several additional problems arise when fitting, post-processing, storing, and interpolating LTCs in the anisotropic case. Each of these operations must be done carefully to avoid rendering artifacts. We find robust solutions for each operation by introducing and exploiting invariance properties of LTCs. As a result, we obtain a small 84 look-up table that provides a plausible and artifact-free LTC approximation to anisotropic GGX and brings it to real-time area-light shading.

Abstract

Precomputed Radiance Transfer (PRT) can achieve high quality renders of glossy materials at real-time framerates. PRT involves precomputing a 𝑘-dimensional transfer vector of Spherical Harmonic (SH) coefficients at specific points for a scene. Most prior art precomputes transfer at vertices of the mesh and interpolates color for interior points. They require finer mesh tessellations for high quality renderings. In this paper, we explore and present the use of textures for storing transfer. Using transfer textures decouples mesh resolution from transfer storage and sampling which is useful especially for glossy renders. We further demonstrate glossy interreflections by precomputing additional textures. We thoroughly discuss practical aspects of transfer textures and analyze their performance in realtime rendering applications. We show equivalent or higher render quality and FPS and demonstrate results on several challenging scenes

Abstract

Histopathology image analysis is widely accepted as a gold standard for cancer diagnosis. The Cancer Genome Atlas (TCGA) contains large repositories of histopathology whole slide images spanning several organs and subtypes. However, not much work has gone into analyzing all the organs and subtypes and their similarities. Our work attempts to bridge this gap by training deep learning models to classify cancer vs. normal patches for 11 subtypes spanning seven organs (9,792 tissue slides) to achieve high classification performance. We used these models to investigate their performances in the test set of other organs (cross-organ inference). We found that every model had a good cross-organ inference accuracy when tested on breast, colorectal, and liver cancers. Further, high accuracy is observed between models trained on the cancer subtypes originating from the same organ (kidney and lung). We also validated these performances by showing the separability of cancer and normal samples in a high-dimensional feature space. We further hypothesized that the high cross-organ inferences are due to shared tumor morphologies among organs. We validated the hypothesis by showing the overlap in the Gradient-weighted Class Activation Mapping (GradCAM) visualizations and similarities in the distributions of nuclei features present within the high-attention regions.

Abstract

Reading Indian scene texts is complex due to the use of regional vocabulary, multiple fonts/scripts, and text size. This work investigates the significant differences in Indian and Latin Scene Text Recognition (STR) systems. Recent STR works rely on synthetic generators that involve diverse fonts to ensure robust reading solutions. We present utilizing additional non-Unicode fonts with generally employed Unicode fonts to cover font diversity in such synthesizers for Indian languages. We also perform experiments on transfer learning among six different Indian languages. Our transfer learning experiments on synthetic images with common backgrounds provide an exciting insight that Indian scripts can benefit from each other than from the extensive English datasets. Our evaluations for the real settings help us achieve significant improvements over previous methods on four Indian languages from standard datasets like IIIT-ILST, MLT-17, and the new dataset (we release) containing 440 scene images with 500 Gujarati and 2535 Tamil words. Further enriching the synthetic dataset with non-Unicode fonts and multiple augmentations helps us achieve a remarkable Word Recognition Rate gain of over 33% on the IIIT-ILST Hindi dataset. We also present the results of lexicon-based transcription approaches for all six languages

Abstract

In many Asian countries with unconstrained road traffic conditions, driving violations such as not wearing helmets and triple-riding are a significant source of fatalities involving motorcycles. Identifying and penalizing such riders is vital in curbing road accidents and improving citizens’ safety. With this motivation, we propose an approach for detecting, tracking, and counting motorcycle riding violations in videos taken from a vehicle-mounted dashboard camera. We employ a curriculum learning-based object detector to better tackle challenging scenarios such as occlusions. We introduce a novel trapezium-shaped object boundary representation to increase robustness and tackle the rider-motorcycle association. We also introduce an amodal regressor that generates bounding boxes for the occluded riders. Experimental results on a large-scale unconstrained driving dataset demonstrate the superiority of our approach compared to existing approaches and other ablative variants.

Abstract

The variability of clinical course and prognosis of COVID-19 highlights the necessity of patient sub-group risk stratification based on clinical data. In this study, clinical data from a cohort of Indian COVID-19 hospitalized patients is used to develop risk stratification and mortality prediction models. We analyzed a set of 70 clinical parameters including physiological and hematological for developing machine learning models to identify biomarkers. We also compared the Indian and Wuhan cohort, and analyzed the role of steroids. A bootstrap averaged ensemble of Bayesian networks was also learned to construct an explainable model for discovering actionable influences on mortality and days to outcome. We discovered blood parameters, diabetes, co-morbidity and SpO2 levels as important risk stratification features, whereas mortality prediction is dependent only on blood parameters. XGboost and logistic regression model yielded the best performance on risk stratification and mortality prediction, respectively (AUC score 0.83, AUC score 0.92). Blood coagulation parameters (ferritin, D-Dimer and INR), immune and inflammation parameters IL6, LDH and Neutrophil (%) are common features for both risk and mortality prediction. Compared with Wuhan patients, Indian patients with extreme blood parameters indicated higher survival rate. Analyses of medications suggest that a higher proportion of survivors and mild patients who were administered steroids had extreme neutrophil and lymphocyte percentages. The ensemble averaged Bayesian network structure revealed serum ferritin to be the mos

Abstract

Many populous countries including India areburdened with a considerable backlog of legalcases. Development of automated systems thatcould process legal documents and augmentlegal practitioners can mitigate this. However,there is a dearth of high-quality corpora that isneeded to develop such data-driven systems.The problem gets even more pronounced inthe case of low resource languages such asHindi. In this resource paper, we introducetheHindi Legal Documents Corpus (HLDC), acorpus of more than 900K legal documents inHindi. Documents are cleaned and structuredto enable the development of downstream ap-plications. Further, as a use-case for the cor-pus, we introduce the task of bail prediction.We experiment with a battery of models andpropose a Multi-Task Learning (MTL) basedmodel for the same. MTL models use sum-marization as an auxiliary task along with bailprediction as the main task. Experiments withdifferent models are indicative of the needfor further research in this area. We releasethe corpus and model implementation codewith this paper:https://github.com/Exploration-Lab/HLDC

Abstract

Quadrotors find a vast potential use in delivery and disaster relief operations. Control becomes critical in such sce- narios, especially when quadrotors have to manoeuvre through constrained spaces or deliver payloads at precise locations in the presence of external disturbances and parametric uncertainties stemming from uncertain payloads. Therefore, the controller has to guarantee a predefined tracking accuracy not to violate the state constraints. On the other hand, conventional fixed-valued state constraints are not suitable in many scenarios such as (i) initial offset being well beyond the expected accuracy, (ii) system dynam- ics experiencing significant transients due to the dropping of the payload. However, to the best of the authors’ knowledge, state- of-the-art controllers do not provide any solution for an underac- tuated system like a quadrotor when the system needs to honour time-varying constraints under uncertainties. This work proposes a controller for quadrotors which is robust against external distur- bances and parametric variations and guarantees a time-varying predefined position, velocity, attitude, and attitude-rate accuracy. The closed-loop system stability is established analytically, and the effectiveness of the proposed controller is validated experimentally compared to the state-of-the-art under a precision payload delivery scenario.

Abstract

Effective design of autopilots for fixed-wing unmanned aerial vehicles (UAVs) is still a great challenge, due to unmodeled effects and uncertainties that these vehi- cles exhibit during flight. Unmodeled effects and uncertain- ties comprise longitudinal/lateral cross-couplings, as well as poor knowledge of equilibrium points (trimming points) of the UAV dynamics. The main contribution of this article is a new adaptive autopilot design, based on uncertain Euler–Lagrange dynamics of the UAV and where the control can explicitly take into account under-actuation in the dy- namics, reduced structural knowledge of cross-couplings and trimming points. This system uncertainty is handled via appropriately designed adaptive laws: stability of the con- trolled UAV is analyzed. Hardware-in-the-loop tests, com- parisons with an Ardupilot autopilot and with a robustified autopilot validate the effectiveness of the control design, even in the presence of strong saturation of the UAV actua- tors.