Yaroslav Dokuchaev·6d agoMedical imaging has solved the problem of sharing images.In a new article published in The European Business Review, we explore why DICOM, despite becoming the foundation of modern medical…
Büşra KARAKILINÇ·Sep 21DICOM Data Pipeline: Reading, EDA, and PreprocessingAn end-to-end build-out: reading DICOM-format chest X-rays in Python, running exploratory data analysis on 26,684 images, constructing a…
Najm F.·Aug 30Your Segmentation Model Shipped. The Case Still Isn’t Done.A mask is an intermediate file. The clinical job is finishing the workflow.
Najm F.·Aug 30Your CT Viewer Didn’t Crash. Chrome Ran Out of Patience.The demo used a polite study. Production sent the rude one. Worker, WASM, and WebGL are how you survive.
Najm F.·Aug 22The Imaging Demo Is Not the Production SystemOne user, one GPU, one afternoon. Then concurrent sessions, RAM bleed, and data that is not allowed to leave the country.
Najm F.·Aug 22A DICOM Viewer Is Not a Clinical ProductYou can embed a generic viewer in a week. Then a pathologist, an orthodontist, or an AI vendor asks for the actual job.
Najm F.·Aug 20DICOMweb Is Not “DICOM but REST” (Even If That’s How Everyone Sells It)QIDO, WADO, and STOW make imaging look modern — until caching, auth, and partial studies show up
Najm F.·Aug 20The PHI Hiding Inside the PixelsYou cleaned the DICOM tags. Great. The patient name may still be burned into the image.
Najm F.·Aug 18PACS Is Not Just a Hard Drive for Medical ImagesIf you think PACS only stores scans, you are missing the workflow machine underneath
Najm F.·Aug 14Your AI doesn’t fail because of the model. It fails because of the DICOM.Clean research datasets lie. Hospital imaging is messy, incomplete, duplicated, mislabeled — and that’s where models fall apart
