Bio

Dr. Jeya Maria Jose Valanarasu is a postdoctoral scholar working with the Stanford Machine Learning Group and the Center for Artificial Intelligence in Medicine and Imaging (AIMI Center). He leads the AI for Healthcare bootcamp with Dr. Andrew Ng, Dr. Curt Langlotz, and Dr. Nigam Shah which provides Stanford students an opportunity to engage in advanced research at the intersection of AI and healthcare.

He obtained his Ph.D. and M.S from Johns Hopkins University where he worked on various problems in Computer Vision, Machine Learning, and Healthcare. His research aims to overcome the challenges that arise when translating machine learning models to practical applications for healthcare and engineering sectors. His works have spanned over topics like designing effective deep architectures, model adaptability to changing environments, role of data and annotations, multi-modal learning and taming large models for computer vision and healthcare tasks. He has published over 25 peer-reviewed journal/conference articles at top venues and filed 3 U.S. patents. He has been awarded Amazon Research Fellowship 2022, Best Student Paper Awards at ICRA 2022, CVIP 2019, MICCAI Young Scientist Impact Award Finalist 2022, and the NIH MICCAI Award 2022. He has also served as a reviewer for multiple journals and conferences.

Honors & Awards

  • Amazon Research Fellowship, Amazon (2022)
  • Young Scientist Impact Award - Finalist, MICCAI (2022)
  • NIH MICCAI Award, NIH (2022)
  • Outstanding Paper Award, ICRA (2021)
  • Best Student Paper Award, CVIP (2019)

Professional Education

  • Doctor of Philosophy, Johns Hopkins University (2023)
  • Master of Science, Johns Hopkins University (2021)
  • Bachelor of Technology, National Institute of Technology (2019)

Stanford Advisors

Contact

  • Academic
    jmjose@stanford.edu
    University - Scholar Department: Radiology Position: Postdoctoral Scholar

Additional Info

  • Mail Code: 5128

Links

Lab Affiliations

All Publications

  • Fine-Context Shadow Detection using Shadow Removal Valanarasu, J., Patel, V. M., IEEE IEEE COMPUTER SOC. 2023: 1705-1714

    View details for DOI 10.1109/WACV56688.2023.00175

    View details for Web of Science ID 000971500201078

  • KiU-Net: Overcomplete Convolutional Architectures for Biomedical Image and Volumetric Segmentation IEEE TRANSACTIONS ON MEDICAL IMAGING Valanarasu, J., Sindagi, V. A., Hacihaliloglu, I., Patel, V. M. 2022; 41 (4): 965-976

    Abstract

    Most methods for medical image segmentation use U-Net or its variants as they have been successful in most of the applications. After a detailed analysis of these "traditional" encoder-decoder based approaches, we observed that they perform poorly in detecting smaller structures and are unable to segment boundary regions precisely. This issue can be attributed to the increase in receptive field size as we go deeper into the encoder. The extra focus on learning high level features causes U-Net based approaches to learn less information about low-level features which are crucial for detecting small structures. To overcome this issue, we propose using an overcomplete convolutional architecture where we project the input image into a higher dimension such that we constrain the receptive field from increasing in the deep layers of the network. We design a new architecture for im- age segmentation- KiU-Net which has two branches: (1) an overcomplete convolutional network Kite-Net which learns to capture fine details and accurate edges of the input, and (2) U-Net which learns high level features. Furthermore, we also propose KiU-Net 3D which is a 3D convolutional architecture for volumetric segmentation. We perform a detailed study of KiU-Net by performing experiments on five different datasets covering various image modalities. We achieve a good performance with an additional benefit of fewer parameters and faster convergence. We also demonstrate that the extensions of KiU-Net based on residual blocks and dense blocks result in further performance improvements. Code: https://github.com/jeya-maria-jose/KiU-Net-pytorch.

    View details for DOI 10.1109/TMI.2021.3130469

    View details for Web of Science ID 000777332500020

    View details for PubMedID 34813472

  • TRANSFORMER-BASED SAR IMAGE DESPECKLING Perera, M. V., Bandara, W., Valanarasu, J., Patel, V. M., IEEE IEEE. 2022: 751-754

    View details for DOI 10.1109/IGARSS46834.2022.9884596

    View details for Web of Science ID 000920916600185

  • Hyperspectral Pansharpening Based on Improved Deep Image Prior and Residual Reconstruction IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING Bandara, W., Valanarasu, J., Patel, V. M. 2022; 60

    View details for DOI 10.1109/TGRS.2021.3139292

    View details for Web of Science ID 000764883100023

  • TransWeather: Transformer-based Restoration of Images Degraded by Adverse Weather Conditions Valanarasu, J., Yasarla, R., Patel, V. M., IEEE COMP SOC IEEE COMPUTER SOC. 2022: 2343-2353

    View details for DOI 10.1109/CVPR52688.2022.00239

    View details for Web of Science ID 000867754202060

  • Simultaneous Bone and Shadow Segmentation Network Using Task Correspondence Consistency Rahman, A., Valanarasu, J., Hacihaliloglu, I., Patel, V. M., Wang, L., Dou, Q., Fletcher, P. T., Speidel, S., Li, S. SPRINGER INTERNATIONAL PUBLISHING AG. 2022: 330-339

    View details for DOI 10.1007/978-3-031-16440-8_32

    View details for Web of Science ID 000867306400032

  • UNeXt: MLP-Based Rapid Medical Image Segmentation Network Valanarasu, J., Patel, V. M., Wang, L., Dou, Q., Fletcher, P. T., Speidel, S., Li, S. SPRINGER INTERNATIONAL PUBLISHING AG. 2022: 23-33

    View details for DOI 10.1007/978-3-031-16443-9_3

    View details for Web of Science ID 000867370700003

  • Orientation-Guided Graph Convolutional Network for Bone Surface Segmentation Rahman, A., Bandara, W., Valanarasu, J., Hacihaliloglu, I., Patel, V. M., Wang, L., Dou, Q., Fletcher, P. T., Speidel, S., Li, S. SPRINGER INTERNATIONAL PUBLISHING AG. 2022: 412-421

    View details for DOI 10.1007/978-3-031-16443-9_40

    View details for Web of Science ID 000867370700040

  • SPIN Road Mapper: Extracting Roads from Aerial Images via Spatial and Interaction Space Graph Reasoning for Autonomous Driving Bandara, W., Valanarasu, J., Patel, V. M., IEEE IEEE. 2022

    View details for DOI 10.1109/ICRA.46639.2022.9812134

    View details for Web of Science ID 000941265700033

  • SAR DESPECKLING USING OVERCOMPLETE CONVOLUTIONAL NETWORKS Perera, M. V., Bandara, W., Valanarasu, J., Patel, V. M., IEEE IEEE. 2022: 401-404

    View details for DOI 10.1109/IGARSS46834.2022.9884632

    View details for Web of Science ID 000920916600099

  • Exploring Overcomplete Representations for Single Image Deraining Using CNNs IEEE JOURNAL OF SELECTED TOPICS IN SIGNAL PROCESSING Yasarla, R., Valanarasu, J., Patel, V. M. 2021; 15 (2): 229-239

    View details for DOI 10.1109/JSTSP.2020.3039393

    View details for Web of Science ID 000622098600007

  • OVERCOMPLETE REPRESENTATIONS AGAINST ADVERSARIAL VIDEOS Lo, S., Valanarasu, J., Patel, V. M., IEEE IEEE. 2021: 1939-1943

    View details for DOI 10.1109/ICIP42928.2021.9506537

    View details for Web of Science ID 000819455102013

  • Overcomplete Deep Subspace Clustering Networks Valanarasu, J., Patel, V. M., IEEE IEEE COMPUTER SOC. 2021: 746-755

    View details for DOI 10.1109/WACV48630.2021.00079

    View details for Web of Science ID 000692171000075

  • Over-and-Under Complete Convolutional RNN for MRI Reconstruction Guo, P., Valanarasu, J., Wang, P., Zhou, J., Jiang, S., Patel, V. M., deBruijne, M., Cattin, P. C., Cotin, S., Padoy, N., Speidel, S., Zheng, Y., Essert, C. SPRINGER INTERNATIONAL PUBLISHING AG. 2021: 13-23

    Abstract

    Reconstructing magnetic resonance (MR) images from under-sampled data is a challenging problem due to various artifacts introduced by the under-sampling operation. Recent deep learning-based methods for MR image reconstruction usually leverage a generic auto-encoder architecture which captures low-level features at the initial layers and high-level features at the deeper layers. Such networks focus much on global features which may not be optimal to reconstruct the fully-sampled image. In this paper, we propose an Over-and-Under Complete Convolutional Recurrent Neural Network (OUCR), which consists of an overcomplete and an undercomplete Convolutional Recurrent Neural Network (CRNN). The overcomplete branch gives special attention in learning local structures by restraining the receptive field of the network. Combining it with the undercomplete branch leads to a network which focuses more on low-level features without losing out on the global structures. Extensive experiments on two datasets demonstrate that the proposed method achieves significant improvements over the compressed sensing and popular deep learning-based methods with less number of trainable parameters.

    View details for DOI 10.1007/978-3-030-87231-1_2

    View details for Web of Science ID 000712022300002

    View details for PubMedID 34661201

    View details for PubMedCentralID PMC8517933

  • Medical Transformer: Gated Axial-Attention for Medical Image Segmentation Valanarasu, J., Oza, P., Hacihaliloglu, I., Patel, V. M., DeBruijne, M., Cattin, P. C., Cotin, S., Padoy, N., Speidel, S., Zheng, Y., Essert, C. SPRINGER INTERNATIONAL PUBLISHING AG. 2021: 36-46

    View details for DOI 10.1007/978-3-030-87193-2_4

    View details for Web of Science ID 000712019600004

  • Learning to Segment Brain Anatomy From 2D Ultrasound With Less Data IEEE JOURNAL OF SELECTED TOPICS IN SIGNAL PROCESSING Jose Valanarasu, J., Yasarla, R., Wang, P., Hacihaliloglu, I., Patel, V. M. 2020; 14 (6): 1221-1234

    View details for DOI 10.1109/JSTSP.2020.3001513

    View details for Web of Science ID 000575014600014