Radhika Dua

Radhika Dua

Auto-generated profile, based on publicly available information

Radhika Dua

KAIST

Research Focus & Metrics

Academic Career Length

14Years (2013-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 August 2013 in Pediatric Emergency Care
Verified
Journal Article
Closed access
SFShahme Ahamed Farook
Shahme Ahamed Farook
ADAnika Kim Jap Davis
Anika Kim Jap Davis
ZSZaid Sadiq
Zaid Sadiq
RDRadhika Dua
Radhika Dua
LNLaurence Newman
Laurence Newman
A study was carried out to determine the influence of telemedicine on management of facial lacerations in children. Proforma was designed, and information was collected retrospectively over a 4-month period. Of the 2524 telemedicine referrals, 27% (693/2524) were pediatric referrals. One hundred forty-five of these were children with facial lacerations, of which only 98 were admitted for possible surgical management. During the same period, a further 45 children were admitted to the pediatric ward through verbal referral. The average age of the children admitted was 4.85 years and male-to-female ratio was 2:1. Lacerations to the lips (42%) were the most common site of injury, and the majority sustained the injury as a result of a mechanical fall (42%). Among the telemedicine group, 86% had surgery, whereas in the nontelemedicine group it was 82%. On average, 2 to 3 photographs accompanied the referrals, and none of the patients who were treated conservatively by the referring centers following telemedicine referral and discussion with Queen Victoria Hospital required a re-referral. The results from this study confirm that clinical findings together with images from telemedicine provide adequate information to make clinical decisions in children with facial lacerations.

1 June 2020 in Chest
Verified
Journal Article
Open access
PDF file
AKA. Kohli
A. Kohli
RDR. Dua
R. Dua
VHVan Holden
Van Holden
EPE. Pickering
E. Pickering
ASAshutosh Sachdeva
Ashutosh Sachdeva
PCP. Costantini
P. Costantini
TYPE: Abstract Publication TOPIC: Procedures PURPOSE: Foreign Body Aspiration (FBA) is uncommon in adults and compared to children, diagnosis of FBA is significantly delayed in adults. We present a case of a woman presenting with hemoptysis, diagnosed with iron pill aspiration (IPA). METHODS: A 67-year old woman with a history of COPD presented with persistent cough and hemoptysis. Imaging studies including CT chest failed to reveal a clear etiology. She underwent bronchoscopy, which revealed a necrotic lesion in RLL. Biopsy revealed abundant refractory material, which stained strongly with prussian blue for iron. RESULTS: Non-asphyxiating FBA in adults present with cough, hemoptysis or dyspnea. Iron deposition leads to bronchial wall inflammation, acute mucosal damage, granuloma formation and fibrosis leading to stenosis. IPA is usually diagnosed with a triad of history of pill aspiration, airway inflammation and iron particles in bronchial biopsy without the discovery of the pill. Endobronchial biopsy may show iron deposits even a year after aspiration. CONCLUSIONS: Pill aspiration is oftentimes not recognized as a cause of chronic symptoms. IPA should be considered in patients with impaired swallowing, who present with hemoptysis with unimpressive imaging. Importance of careful history-taking and review of patient’s medications is highlighted. CLINICAL IMPLICATIONS: Iron tablets are commonly prescribed to patients and are left on a patient’s medication list long after their iron stores are replete. A relatively “benign” medication may cause severe side effects. This emphasises the importance of medicine reconciliation and keeping high index of suspicion of FBA in patients with persistent cough/hemoptysis. DISCLOSURE: No significant relationships. KEYWORDS: iron pill aspiration, hemoptysis, foreign body aspiration

14 April 2022 in arXiv
Verified
Preprint
Open access
PDF file
RSRahul Sajnani
Rahul Sajnani
APAdrien Poulenard
Adrien Poulenard
JJJivitesh Jain
Jivitesh Jain
RDRadhika Dua
Radhika Dua
LGLeonidas J. Guibas
Leonidas J. Guibas
SSSrinath Sridhar
Srinath Sridhar
Progress in 3D object understanding has relied on manually canonicalized shape datasets that contain instances with consistent position and orientation (3D pose). This has made it hard to generalize these methods to in-the-wild shapes, eg., from internet model collections or depth sensors. ConDor is a self-supervised method that learns to Canonicalize the 3D orientation and position for full and partial 3D point clouds. We build on top of Tensor Field Networks (TFNs), a class of permutation- and rotation-equivariant, and translation-invariant 3D networks. During inference, our method takes an unseen full or partial 3D point cloud at an arbitrary pose and outputs an equivariant canonical pose. During training, this network uses self-supervision losses to learn the canonical pose from an un-canonicalized collection of full and partial 3D point clouds. ConDor can also learn to consistently co-segment object parts without any supervision. Extensive quantitative results on four new metrics show that our approach outperforms existing methods while enabling new applications such as operation on depth images and annotation transfer.

26 July 2022 in arXiv
Verified
Preprint
Open access
PDF file
RDRadhika Dua
Radhika Dua
SYSeongjun Yang
Seongjun Yang
YLYixuan Li
Yixuan Li
ECEdward Choi
Edward Choi
Despite the recent advances in out-of-distribution(OOD) detection, anomaly detection, and uncertainty estimation tasks, there do not exist a task-agnostic and post-hoc approach. To address this limitation, we design a novel clustering-based ensembling method, called Task Agnostic and Post-hoc Unseen Distribution Detection (TAPUDD) that utilizes the features extracted from the model trained on a specific task. Explicitly, it comprises of TAP-Mahalanobis, which clusters the training datasets' features and determines the minimum Mahalanobis distance of the test sample from all clusters. Further, we propose the Ensembling module that aggregates the computation of iterative TAP-Mahalanobis for a different number of clusters to provide reliable and efficient cluster computation. Through extensive experiments on synthetic and real-world datasets, we observe that our approach can detect unseen samples effectively across diverse tasks and performs better or on-par with the existing baselines. To this end, we eliminate the necessity of determining the optimal value of the number of clusters and demonstrate that our method is more viable for large-scale classification tasks.

8 August 2026
Verified
Conference Paper
Open access
RDRadhika Dua
Radhika Dua
AAAditi Agarwal
Aditi Agarwal
AJAishwarya Jayagopal
Aishwarya Jayagopal
DSDepanshu Sani
Depanshu Sani
AWAlex Wilson
Alex Wilson
VRVaibhav Rajan
Vaibhav Rajan