Dr. Harshita Sharma is a Senior ML Researcher in Health Intelligence, Microsoft Research Cambridge. She works in the Medical Imaging team, where she explores machine learning solutions aimed towards improving patient outcomes and clinical workflows. Harshita’s research interests include machine learning, image analysis, computer vision, and multimodal clinical data analysis for healthcare.
Until June 2021, Harshita worked as a PostDoc in the Department of Engineering Science at the University of Oxford, and was a member of the Biomedical Image Analysis (BioMedIA) group, advised by Prof. Alison Noble. She worked in the interdisciplinary ERC Project PULSE (Perception Ultrasound by Learning Sonographic Experience) with the aim to develop multi-modal analysis methods to model sonographic experience and make obstetric ultrasound imaging easier for non-specialists.
In 2017, Harshita obtained her Ph.D. (Dr.-Ing.) in Computer Vision from Technische Universität Berlin advised by Prof. Olaf Hellwich, after being awarded a full-time DAAD Ph.D. Scholarship (2013-17). She was a member of the Computer Vision and Remote Sensing group (CVRS) at TU Berlin. She collaborated with Charité University Hospital, VMScope GmbH Berlin, and University Hospital Schleswig-Holstein. She analysed histopathological whole slide images of gastric carcinoma to develop methods in computational pathology for automatic cancer grade classification, necrosis detection and cell nuclei classification.
In 2012-13, Harshita was a lecturer in electronics and communication engineering at the Jaypee Institute of Information Technology. In 2011-12, she was fortunate to be an exchange research scholar at Technical University Berlin under the DAAD-IIT Master Sandwich Programme, where she explored histopathological breast cancer biopsies for content-based image retrieval and collaborated in the Virtual Specimen Scout project of Charité University Hospital Berlin, TU Berlin, VMscope GmbH Berlin and NEXUS AG Villingen-Schwenningen. In 2012, Harshita obtained her Masters degree (M.Tech.) in electrical engineering from the Indian Institute of Technology Roorkee, and in 2010, and Bachelors degree (B.Tech.) in electronics and communication engineering from the Indira Gandhi Delhi Technical University of Women.
Ph.D.(Dr.-Ing.) in Computer Vision, 2017
Technische Universität Berlin
M.Tech. in Instrumentation and Signal Processing, 2012
Indian Institute of Technology Roorkee
B.Tech. in Electronics and Communication Engineering, 2010
Indira Gandhi Delhi Technical University for Women
IEEE TPAMI, IEEE TMI, Nature Comms, PLOS ONE, IEEE JBHI, IEEE RA-L, IEEE Access, CMIG (Elsevier), AIIM (Elsevier), CPMB (Elsevier), IJCARS (Springer), PeerJ, Ultrasonic Imaging, Journal of Pathology, MDPI J Imaging, MDPI Applied Science, Biomedical Optics Express, International Journal of Imaging Systems and Technology (IMA)-Editorial Board Member, Oxford Initiatives.
ISBI 2022, MICCAI-ASMUS 2021 (Programme Committee Member), MIUA 2021 (Session Chair), MICCAI 2021, ETRA 2021, IPMI 2021, IEEE ISBI 2021, ECDP 2021 (Scientific Committee Member and Moderator), MICCAI-ASMUS 2020, MICCAI 2020, MIUA 2020, MIDL 2020, ISBI 2020, MICCAI 2019, MIDL 2019, IEEE ISBI 2019, IEEE ISBI 2018, BioRob 2018, IEEE BIBE 2017 (Programme Committee Member).
UKRI ECR Forum, International Society of Ultrasound in Obstetrics and Gynecology (ISUOG), ESDIP Scientific Committee Member, TU Berlin-Alumni Network (International Contact- GB) and Manager of the LinkedIn TU Berlin Alumni Club UK, MedIAN-Medical Image Analysis Network, Lindau Alumni Network, Heidelberg Laureate Forum - Alumni Network, DAAD Alumni Network, Durtti-The Artificial Intelligence Group, Cancer Epigenetics Society, Science Advisory Board
April 2021 | Harshita gave an invited talk “Artificial Intelligence in Medical Imaging - Methods, Applications, and Challenges” at Microsoft Research India.
April 2021 | Harshita presented her research paper “Multi-modal Learning from Video, Eye Tracking, and Pupillometry for Operator Skill Characterization in Clinical Fetal Ultrasound” as oral presentation at IEEE ISBI 2021.
October 2020 | Harshita presented a hot abstract in ultrasound imaging “Task‐evoked pupillary response as an index of cognitive workload of sonologists undertaking fetal ultrasound”, and equally contributed to the abstract “The data science of obstetric ultrasound: automatic analysis of full‐length anomaly scans using machine learning algorithms” at the ISUOG Virtual World Congress 2020.
October 2020 | A co-authored paper from the PULSE project “A Curriculum Learning Based Approach to Captioning Ultrasound Images” is the Best Presentation Winner at Advances in Simplifying Medical UltraSound-ASMUS 2020 Workshop at MICCAI 2020.
April 2020 | A co-authored paper from the PULSE project “Discovering Salient Anatomical Landmarks by Predicting Human Gaze” has won the 2nd runner up for Best Paper Award at ISBI 2020. News items can be found here and here.
July 2019 | A co-authored paper from the PULSE project “Towards Capturing Sonographic Experience: Cognition-Inspired Ultrasound Video Saliency Prediction” has won the Best Paper Award at MIUA 2019.
April 2019 | Harshita presented her Postdoctoral research paper “Spatio-temporal partitioning and description of full-length routine fetal anomaly ultrasound scans” as an oral presentation in IEEE ISBI 2019 at Venice, Italy.
March 2018 | Harshita presented her research at the Oxbridge Women in Computer Science Conference 2018 at Cambridge, UK.
June 2018 | Harshita participated in the 68th Lindau Nobel Laureate Meeting 2018 as a Mars Fellow at Lindau, Germany. She has shared her experience here!
May 2018 | Hashita has a feature in the TU Berlin Newsletter TU Intern!
March 2018 | Harshita’s interview (invited feature) by Durtti.com-The Artificial Intelligence Group: Research Scientist Dr. Harshita Sharma Talks To Durtti About The Future Of Biomedical Imaging.
April 2017 | Harshita defended her PhD thesis at TU Berlin and received the Dr.-Ing. degree 😄.
March 2017 | Harshita gave an invited talk Medical Image Analysis using Modern Description and Learning Techniques at NIEC, Delhi.
October 2016 | Harshita presented her PhD research at the workshop Computational Life Sciences @ Bayer in Berlin, Germany.
May 2016 | Harshita’s deep learning in digital pathology talk at ECDP 2016 in Berlin was mentioned in the MicroDimensions Blog Article.
March 2016 | Harshita received student paper and poster awards at the SPIE Medical Imaging Symposium 2016 in San Diego, United States. Find out more here!
January 2016 | Harshita’s PhD research is featured in the Computer Vision & Remote Sensing Group Activities of the TU Berlin.
August 2015 | Harshita presented her PhD research and participated in the 3rd Heidelberg Laureate Forum at Heidelberg, Germany!
June 2014 | Harshita gave an invited talk Histopathogical Image Analysis for Content-Based Image Retrieval and Classification at the DAAD WISE scholarship holders’ meeting at the Frei Universität in Berlin, Germany.