Amita Kamath

ORCID: 0000-0002-5005-415X
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Hepatocellular Carcinoma Treatment and Prognosis
  • Liver Disease Diagnosis and Treatment
  • MRI in cancer diagnosis
  • Pelvic floor disorders treatments
  • Multimodal Machine Learning Applications
  • Diverticular Disease and Complications
  • Maternal and fetal healthcare
  • Ectopic Pregnancy Diagnosis and Management
  • Domain Adaptation and Few-Shot Learning
  • Radiomics and Machine Learning in Medical Imaging
  • Advanced Image and Video Retrieval Techniques
  • Anorectal Disease Treatments and Outcomes
  • Pregnancy and preeclampsia studies
  • Liver Disease and Transplantation
  • Advanced MRI Techniques and Applications
  • Prenatal Screening and Diagnostics
  • Medical Imaging Techniques and Applications
  • Ovarian cancer diagnosis and treatment
  • Cerebrovascular and Carotid Artery Diseases
  • Cancer Risks and Factors
  • Gastrointestinal Tumor Research and Treatment
  • Language, Metaphor, and Cognition
  • Gastrointestinal disorders and treatments
  • Renal cell carcinoma treatment
  • Renal and related cancers

Icahn School of Medicine at Mount Sinai
2016-2025

Mount Sinai Medical Center
2024

Mount Sinai Hospital
2019-2024

Society of Interventional Radiology
2023

Mount Sinai Hospital
2023

University of California, Los Angeles
2023

Mount Sinai Hospital
2019-2022

Mount Sinai Health System
2018-2019

Texas Medical Association
2016

NYU Langone Health
2010

Severe obstetric hemorrhage is a leading cause of maternal mortality and morbidity worldwide. Major in the antepartum period presents potential risks for both mother fetus. Similarly, postpartum (PPH) accounts up to quarter deaths Potential causes severe that radiologists should be familiar with include placental abruption, placenta previa, accreta spectrum disorders, vasa previa. Common PPH authors discuss uterine atony, puerperal genital hematomas, rupture dehiscence, retained products...

10.1148/rg.230164 article EN Radiographics 2024-03-28

The Pelvic Floor Disorders Consortium (PFDC) is a multidisciplinary organization of colorectal surgeons, urogynecologists, urologists, gynecologists, gastroenterologists, radiologists, physiotherapists, and other advanced care practitioners. Specialists from these fields are all dedicated to the diagnosis management patients with pelvic floor conditions, but they approach, evaluate, treat such their own unique perspectives given differences in respective training. PFDC was formed bridge gaps...

10.2214/ajr.21.26488 article EN American Journal of Roentgenology 2021-09-10

Performant vision-language (VL) models like CLIP represent captions using a single vector. How much information about language is lost in this bottleneck? We first curate CompPrompts, set of increasingly compositional image that VL should be able to capture (e.g., object, object+property, multiple interacting objects). Then, we train text-only recovery probes aim reconstruct from single-vector text representations produced by several models. This approach does not require images, allowing us...

10.18653/v1/2023.emnlp-main.301 article EN cc-by Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing 2023-01-01

The need for liver transplants is increasing because the prevalence of diseases and indications are growing. In response to shortage grafts from deceased donors, more being performed worldwide with living donors. Radiologic evaluation an integral component in assessment donor candidates ensure their eligibility choose most appropriate surgical approach. MRI preferred modality parenchyma biliary tree. centers, a combination CT used take advantage higher spatial resolution arteries. However,...

10.1148/rg.2021210012 article EN Radiographics 2021-10-01

Recent vision-language (VL) models are powerful, but can they reliably distinguish “right” from “left”? We curate three new corpora to quantify model comprehension of such basic spatial relations. These tests isolate reasoning more precisely than existing datasets like VQAv2, e.g., our What’sUp benchmark contains sets photographs varying only the relations objects, keeping their identity fixed (see Figure 1: must comprehend not usual case a dog under table, also, same on top table). evaluate...

10.18653/v1/2023.emnlp-main.568 article EN cc-by Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing 2023-01-01

The objective of our study was to compare diffusion-weighted imaging (DWI) sequences using a bipolar versus monopolar single-shot echo-planar (EPI) gradient design for image quality and lesion detection characterization in patients with liver disease.In this retrospective study, 77 chronic disease who underwent MRI including DWI at 1.5 T were assessed. Two independent observers reviewed the studies lesions. reference standard diagnosis established by consensus review two different...

10.2214/ajr.13.11695 article EN American Journal of Roentgenology 2014-12-24

Reports have indicated an association of large vessel peripheral arterial occlusion in the setting Coronavirus Disease 2019 (COVID-19). While prior investigations mostly focused on venous or cerebral occlusions, we examined patients presenting exclusively with extremity occlusions to investigate for any predisposing factors this subset COVID-19 patients.This is a retrospective study multi-hospital health care system New York City between February 1st, 2020 and April 30th, 2020. Patient data...

10.1016/j.clinimag.2020.11.023 article EN other-oa Clinical Imaging 2020-11-14

Computer vision systems today are primarily N-purpose systems, designed and trained for a predefined set of tasks. Adapting such to new tasks is challenging often requires non-trivial modifications the network architecture (e.g. adding output heads) or training process losses). To reduce time expertise required develop applications, we would like create general purpose that can learn perform range without any modification learning process. In this paper, propose GPV-1, task-agnostic...

10.48550/arxiv.2104.00743 preprint EN other-oa arXiv (Cornell University) 2021-01-01
Coming Soon ...