Additional completed Projects
Miccai (Medical Image Computing and Computer Assisted Intervention Society) 2022
Significantly contributed towards the challenge hosted by Dr. Valentina Pedoia and the Center-of Intelligent Imaging lab on: "K2S: from under sampled K-space to Automatic Segmentation. The challenge focused on the efficient segmentation of key anatomical structures from under-sampled data which can eventually lead to a potential reduction in scan time for screening knee abnormalities in the clinic. My role was to: i) Curate a dataset of 300 high-resolution 3D knee MRIs including raw k-space data and post-processing annotations with masks for tissue segmentation. The 8x under-sampled, multi-channel k-space data of 300 fat-suppressed 3D FSE Cube sequences along with segmentation masks of cartilage, meniscus, and bone were shared as part of the challenge. These segmentation masks were generated by DL algorithms validated previously. Fully sampled images were shared for training as auxiliary information. ii) Curate a similar inference test set of 50 scans, 8x under-sampled k-spaces. This set did not have fully sampled images. Participants were to submit tissue segmentation labels. I evaluated the challenge submissions with reference to the ground truths, by iii) Dice coefficients (image-specific, tissue-specific), iv) cartilage thickness and bone shape, v) comparisons between submitted and ground truth mesh maps (of bone shape and cartilage thickness).
Additional research contributions in other projects in the MQIR-CI2 lab
"Multiparametric MRI of knees in collegiate basketball players: associations with position, morphological defects, and functional deficits" by Cameron Nosrat
"Automatizing bilateral Hip bone segmentation" by Rafeek Thahakoya, Postdoc scholar
"Calibration study to standardize multi-scanner hip relaxation parameters, T1p and T2" by Koren Roach, PostDoc scholar.
"Association of Hip Capsule morphology with Hip pain in patients without another structural correlate" by Johanna Luitjens, MD
"Predicting locations of OA (tibio femoral/patello femoral/mixed/none) from clinical low resolution MRI images" by Erin Argentieri, rotation grad student
"Detection and classification of Hip Labral Tear from clinical low-resolution MRI scans" by Felix Gassert, MD, Johanna Luitjens MD, and Hector Carbajal, grad student.