about

Smriti Joshi

Smriti Joshi

PhD Researcher · Trustworthy AI in Healthcare

I am a PhD Researcher at BCN-AIM, Universitat de Barcelona, funded by the Horizon Europe Project RadioVal. My work focuses on building models and engaging community for advancing Health AI.

Affiliations & Education

2014 – 2018 MSc Physics BITS Pilani, India
2019 – 2021 MSc IPCV (Honors) Erasmus Mundus
2021 – Present PhD Candidate BCN-AIM · Universitat de Barcelona

Challenges

MAMA-MIA Challenge — Co-organizer
MICCAI 2025

Designed and ran an open challenge on generalizable and fair tumor segmentation and pCR prediction using breast MRI. Challenge page →

ODELIA Challenge — Top-2 Solution
MICCAI 2025

Placed Top-2 in the ODELIA multi-center breast MRI challenge, demonstrating strong cross-site generalizability of our domain-adaptive approach.

Leadership

AI Working Group Lead — RadioVal
2021 – Present

Leading the AI research workstream in RadioVal, an EU Horizon Europe project on clinical validation of AI for breast cancer treatment response.

MICCAI 2024 LMIC Initiative — Committee Member
2024

Served on the Grant Selection & Sponsorship Committee for the MICCAI 2024 LMIC Initiative, leading selection of 50+ early-career researchers from lower-income countries.

International AI Schools — AFRICAI
2021 – Present

Active member of AFRICAI. Coordinated international AI schools across Africa and South Asia, connecting researchers to global medical AI communities.

Impact

Test-Time Adaptation for Breast MRI
MICCAI 2025 — First Author

Robust tumor segmentation under domain shift without retraining — addressing a core bottleneck for real-world deployment of clinical AI across diverse acquisition conditions.

Uncertainty Estimation for Robust AI
SPIE 2024 — First Author

Investigated well-calibrated uncertainty signals for flagging unreliable predictions in breast MRI segmentation — a key ingredient for building clinician trust in AI.

3 Co-authored Journal & Book Publications
2025

JMI review on generative AI for imaging simulations, a Scientific Data paper on large-scale breast MRI datasets, and a book chapter on stakeholder perspectives in AI for healthcare.

medigan — Clinical AI Framework
JMI 2023

Co-authored medigan, an open-source framework for medical image synthesis adopted across multiple research groups for synthetic data generation and augmentation.

Publications

2025

  1. Single Image Test-Time Adaptation via Multi-View Co-Training
    Smriti Joshi, Richard Osuala, Lidia Garrucho, and 4 more authors
    In International Conference on Medical Image Computing and Computer Assisted Intervention, 2025
  2. Simulating dynamic tumor contrast enhancement in breast MRI using conditional generative adversarial networks
    Richard Osuala, Smriti Joshi, Apostolia Tsirikoglou, and 6 more authors
    Journal of Medical Imaging, 2025
  3. A large-scale multicenter breast cancer DCE-MRI benchmark dataset with expert segmentations
    Lidia Garrucho, Kaisar Kushibar, Claire-Anne Reidel, and 3 more authors
    Scientific Data, 2025
  4. Stakeholder Engagement for Trustworthy AI: Experiences and Results from the RadioVal Project on Breast Cancer Imaging
    Miriam Cabrita, Carina Dantas, Harm Akker, and 8 more authors
    In Trustworthy AI in Cancer Imaging Research, 2025

2024

  1. Leveraging epistemic uncertainty to improve tumour segmentation in breast MRI: an exploratory analysis
    Smriti Joshi, Richard Osuala, Lidia Garrucho, and 6 more authors
    In Medical Imaging 2024: Image Processing, 2024

2023

  1. crossmoda_logo.png
    CrossMoDA 2021 challenge: Benchmark of cross-modality domain adaptation techniques for vestibular schwannoma and cochlea segmentation
    Reuben Dorent, Aaron Kujawa, Marina Ivory, and 37 more authors
    Medical Image Analysis, 2023
    Abs
  2. medigan_logo.png
    medigan: a Python library of pretrained generative models for medical image synthesis
    Richard Osuala, Grzegorz Skorupko, Noussair Lazrak, and 8 more authors
    Journal of Medical Imaging, 2023
  3. FUTURE-AI: International consensus guideline for trustworthy and deployable artificial intelligence in healthcare
    Karim Lekadir, and  others
    2023
  4. Pre- to Post-Contrast Breast MRI Synthesis for Enhanced Tumour Segmentation
    Richard Osuala, Smriti Joshi, Apostolia Tsirikoglou, and 4 more authors
    2023
    Code

2021

  1. crossmoda.png
    nn-UNet training on CycleGAN-translated images for cross-modal domain adaptation in biomedical imaging
    Smriti Joshi, Richard Osuala, Carlos Martín-Isla, and 4 more authors
    In International MICCAI Brainlesion Workshop, 2021