Guides, project breakdowns, and lessons learned from building medical imaging platforms — written by the PYCAD team.
Hang a whole fetal ultrasound study into cardiac-gated MIL (FUSE CHD screen): case-level CHD vs healthy, optional critical vs non-critical severity, plus top attention-ranked frames for review. Internal AUC 0.985 / specificity 0.990; CARDIUM 0.513 to 0.944 with CORAL. arXiv:2609.31376.
Hang motion-corrupted DWI stacks into GraphSVR: stack-wise rigid slice-to-volume transforms that cut grid/rotation error ~73% vs FSL eddy under severe motion, strongest on sparse 12/24/32-direction jerky cases. arXiv:2609.24732.
Hang a weak ultrasound lesion overlay from a frozen segmentor, then progressively repair gross miss, FN gaps, and FP blobs with FAPR without retraining the base model. Hard-case BUSI/TN3K average +13.77%. arXiv:2609.18256.
Hang a liver CT and get organ/tumor/vessel overlays, physical volumes and distances from Port B, plus a simulator-shielded resection sequence from VoxelSage. Dual-port: language orchestrates, image math stays off the LLM. arXiv:2609.37648.
Hang a few points or a 2D box on a 3D CT/MR volume and get full-volume anatomy and pathology overlays from SAMI3D-DW. Category-macro Dice 0.8002 after box plus five clicks on 107 categories / 4326 cases. arXiv:2609.25743.
Hang a speckled ultrasound frame and get a despeckled view that keeps fine anatomy from Mask2Restore. Block-wise masked inpainting plus CRCR, no clean target at train time. Simulated SSIM 0.810, PSNR 20.36. arXiv:2609.32844.
We build custom medical imaging platforms — advanced DICOM viewers, AI segmentation, and the clinical systems around them.
Copyright © 2026 PYCAD. All Rights Reserved.