arXiv subjects

Amit Agrawal

Publications and source records attributed to Amit Agrawal.

29 records · Page 2Linked to original sources

arXiv2020

High Brightness Lasing at Sub-micron Enabled by Droop-Free Fin Light-Emitting Diodes

Efficiency droop, i.e., a decline in brightness of LEDs at high electrical currents, has limited the performance of all commercially available LEDs. Until now, it has limited the output power of sub-micron LEDs and lasers to nanowatt range. Here we present a fin p-n junction LED pixel that eliminates efficiency droop, allowing LEDs brightness to increase linearly with injected current. With record current densities of 1000 KA/cm2 (100 mA), the LEDs transition to lasing within the fin, with high brightness. Despite a light extraction efficiency of only 15%, these devices exceed the output power of any previous electrically-driven sub-micron LED or laser pixel by 100 to 1000 times, while showing comparable external quantum efficiencies. Modeling suggests that spreading of the electron-hole recombination region in fin LEDs at high injection levels suppresses the non-radiative Auger recombination processes. Further refinement of this design is expected to enable development of a new generation of high brightness electrically addressable LED and laser pixels for macro- and micro-scale applications.

physics.app-ph↗

arXiv2020

Likelihood of survival of coronavirus in a respiratory droplet deposited on a solid surface

We predict and analyze the drying time of respiratory droplets from a COVID-19 infected subject, which is a crucial time to infect another subject. The drying of the droplet is predicted by diffusion-limited evaporation model for a sessile droplet placed on a partially-wetted surface with a pinned contact line. The variation of droplet volume, contact angle, ambient temperature, and humidity are considered. We analyze the chances of the survival of the viruses present in the droplet, based on the lifetime of the droplets in several conditions, and find that the chances of survival of the virus are strongly affected by each of these parameters. The magnitude of shear stress inside the droplet computed using the model is not large enough to obliterate the virus. We also explore the relationship between the drying time of a droplet and the growth rate of the spread of COVID-19 for five different cities, and find that they are weakly correlated.

cond-mat.soft↗

arXiv2019

Learning to Generate Synthetic Data via Compositing

We present a task-aware approach to synthetic data generation. Our framework employs a trainable synthesizer network that is optimized to produce meaningful training samples by assessing the strengths and weaknesses of a `target' network. The synthesizer and target networks are trained in an adversarial manner wherein each network is updated with a goal to outdo the other. Additionally, we ensure the synthesizer generates realistic data by pairing it with a discriminator trained on real-world images. Further, to make the target classifier invariant to blending artefacts, we introduce these artefacts to background regions of the training images so the target does not over-fit to them. We demonstrate the efficacy of our approach by applying it to different target networks including a classification network on AffNIST, and two object detection networks (SSD, Faster-RCNN) on different datasets. On the AffNIST benchmark, our approach is able to surpass the baseline results with just half the training examples. On the VOC person detection benchmark, we show improvements of up to 2.7% as a result of our data augmentation. Similarly on the GMU detection benchmark, we report a performance boost of 3.5% in mAP over the baseline method, outperforming the previous state of the art approaches by up to 7.5% on specific categories.

cs.CV↗

arXiv2019

Unsupervised 3D Pose Estimation with Geometric Self-Supervision

We present an unsupervised learning approach to recover 3D human pose from 2D skeletal joints extracted from a single image. Our method does not require any multi-view image data, 3D skeletons, correspondences between 2D-3D points, or use previously learned 3D priors during training. A lifting network accepts 2D landmarks as inputs and generates a corresponding 3D skeleton estimate. During training, the recovered 3D skeleton is reprojected on random camera viewpoints to generate new "synthetic" 2D poses. By lifting the synthetic 2D poses back to 3D and re-projecting them in the original camera view, we can define self-consistency loss both in 3D and in 2D. The training can thus be self supervised by exploiting the geometric self-consistency of the lift-reproject-lift process. We show that self-consistency alone is not sufficient to generate realistic skeletons, however adding a 2D pose discriminator enables the lifter to output valid 3D poses. Additionally, to learn from 2D poses "in the wild", we train an unsupervised 2D domain adapter network to allow for an expansion of 2D data. This improves results and demonstrates the usefulness of 2D pose data for unsupervised 3D lifting. Results on Human3.6M dataset for 3D human pose estimation demonstrate that our approach improves upon the previous unsupervised methods by 30% and outperforms many weakly supervised approaches that explicitly use 3D data.

cs.CV↗

arXiv2019

Revisiting the Photon-Drag Effect in Metal Films

The photon-drag effect, the rectified current in a medium induced by conservation of momentum of absorbed or redirected light, is a unique probe of the detailed mechanisms underlying radiation pressure. We revisit this effect in gold, a canonical Drude metal. We discover that the signal for p-polarized illumination in ambient air is affected in both sign and magnitude by adsorbed molecules, opening previous measurements for reinterpretation. Further, we show that the intrinsic sign of the photon-drag effect is contrary to the prevailing intuitive model of direct momentum transfer to free electrons.

physics.optics↗

arXiv2018

"When and Where?": Behavior Dominant Location Forecasting with Micro-blog Streams

The proliferation of smartphones and wearable devices has increased the availability of large amounts of geospatial streams to provide significant automated discovery of knowledge in pervasive environments, but most prominent information related to altering interests have not yet adequately capitalized. In this paper, we provide a novel algorithm to exploit the dynamic fluctuations in user's point-of-interest while forecasting the future place of visit with fine granularity. Our proposed algorithm is based on the dynamic formation of collective personality communities using different languages, opinions, geographical and temporal distributions for finding out optimized equivalent content. We performed extensive empirical experiments involving, real-time streams derived from 0.6 million stream tuples of micro-blog comprising 1945 social person fusion with graph algorithm and feed-forward neural network model as a predictive classification model. Lastly, The framework achieves 62.10% mean average precision on 1,20,000 embeddings on unlabeled users and surprisingly 85.92% increment on the state-of-the-art approach.

cs.SI↗

arXiv2018

CRAFT: Complementary Recommendations Using Adversarial Feature Transformer

Traditional approaches for complementary product recommendations rely on behavioral and non-visual data such as customer co-views or co-buys. However, certain domains such as fashion are primarily visual. We propose a framework that harnesses visual cues in an unsupervised manner to learn the distribution of co-occurring complementary items in real world images. Our model learns a non-linear transformation between the two manifolds of source and target complementary item categories (e.g., tops and bottoms in outfits). Given a large dataset of images containing instances of co-occurring object categories, we train a generative transformer network directly on the feature representation space by casting it as an adversarial optimization problem. Such a conditional generative model can produce multiple novel samples of complementary items (in the feature space) for a given query item. The final recommendations are selected from the closest real world examples to the synthesized complementary features. We apply our framework to the task of recommending complementary tops for a given bottom clothing item. The recommendations made by our system are diverse, and are favored by human experts over the baseline approaches.

cs.CV↗

arXiv2018

Can 3D Pose be Learned from 2D Projections Alone?

3D pose estimation from a single image is a challenging task in computer vision. We present a weakly supervised approach to estimate 3D pose points, given only 2D pose landmarks. Our method does not require correspondences between 2D and 3D points to build explicit 3D priors. We utilize an adversarial framework to impose a prior on the 3D structure, learned solely from their random 2D projections. Given a set of 2D pose landmarks, the generator network hypothesizes their depths to obtain a 3D skeleton. We propose a novel Random Projection layer, which randomly projects the generated 3D skeleton and sends the resulting 2D pose to the discriminator. The discriminator improves by discriminating between the generated poses and pose samples from a real distribution of 2D poses. Training does not require correspondence between the 2D inputs to either the generator or the discriminator. We apply our approach to the task of 3D human pose estimation. Results on Human3.6M dataset demonstrates that our approach outperforms many previous supervised and weakly supervised approaches.

cs.CV↗

arXiv2018

Context Encoding for Semantic Segmentation

Recent work has made significant progress in improving spatial resolution for pixelwise labeling with Fully Convolutional Network (FCN) framework by employing Dilated/Atrous convolution, utilizing multi-scale features and refining boundaries. In this paper, we explore the impact of global contextual information in semantic segmentation by introducing the Context Encoding Module, which captures the semantic context of scenes and selectively highlights class-dependent featuremaps. The proposed Context Encoding Module significantly improves semantic segmentation results with only marginal extra computation cost over FCN. Our approach has achieved new state-of-the-art results 51.7% mIoU on PASCAL-Context, 85.9% mIoU on PASCAL VOC 2012. Our single model achieves a final score of 0.5567 on ADE20K test set, which surpass the winning entry of COCO-Place Challenge in 2017. In addition, we also explore how the Context Encoding Module can improve the feature representation of relatively shallow networks for the image classification on CIFAR-10 dataset. Our 14 layer network has achieved an error rate of 3.45%, which is comparable with state-of-the-art approaches with over 10 times more layers. The source code for the complete system are publicly available.

cs.CV↗

arXiv2016

Surface-Plasmon-Polariton Laser based on an Open-Cavity Fabry-Perot Resonator

Recent years have witnessed growing interest in the development of small-footprint lasers for potential applications in small-volume sensing and on-chip optical communications. Surface-plasmons, electromagnetic modes evanescently confined to metal-dielectric interfaces, offer an effective route to achieving lasing at nanometer-scale dimensions when resonantly amplified in contact with a gain-medium. Here, we achieve visible frequency ultra-narrow linewidth lasing at room-temperature by leveraging surface plasmons propagating in an open Fabry-Perot cavity formed by a flat metal surface coated with a subwavelength-thick layer of optically-pumped gain medium and orthogonally bound by a pair of flat metal sidewalls. Low perturbation transmission-configuration sampling of the lasing plasmon mode is achieved via an evanescently coupled recessed nanoslit, opening the way to high-figure-of-merit refractive-index sensing of analytes interacting with the open cavity.

physics.optics↗

arXiv2011

An Efficient Large-Area Grating Coupler for Surface Plasmon Polaritons

We report the design, fabrication and characterization of a periodic grating of shallow rectangular grooves in a metallic film with the goal of maximizing the coupling efficiency of an extended plane wave (PW) of visible or near-infrared light into a single surface plasmon polariton (SPP) mode on a flat metal surface. A PW-to-SPP power conversion factor > 45 % is demonstrated at a wavelength of 780 nm, which exceeds by an order of magnitude the experimental performance of SPP grating couplers reported to date at any wavelength. Conversion efficiency is maximized by matching the dissipative SPP losses along the grating surface to the local coupling strength. This critical coupling condition is experimentally achieved by tailoring the groove depth and width using a focused ion beam.

physics.optics↗