Medical Files Processing - Visual NLP Demos & Notebooks

Redact PHI from scanned clinical documents with bounding-box precision using the JSL Vision OCR plus Spark NLP DEID pipeline. 100% PHI recall, fully on-prem. (...)

Search medical records by text, image, or both using JSL Vision cross-modal embeddings, with Spark NLP entity extraction and ICD-10, RxNorm and SNOMED coding. (...)

Classify skin lesions (91% across 8 classes) and retrieve visually similar cases with ICD-10 coding, powered by JSL Vision image embeddings. (...)

Classify chest X-ray findings and body parts, extract radiology entities, and retrieve similar prior studies on-prem. (...)

Breast-cancer screening, tissue typing and WBC classification with ICD-O coding over a large histopathology image index. (...)

Extract cardiac measurements (HR, PR, QRS, QT, Axis) from scanned paper ECGs and triage normal vs abnormal with JSL Vision Structured. (...)

Recognize text from DICOM format documents. This feature explores both to the text on the image and to the text from the metadata file. (...)

Deidentify DICOM documents by masking PHI information on the image and by either masking or obfuscating PHI from the metadata. (...)

De-identification of Whole Slide Images (...)

De-identify DICOM files to mask or obfuscate PHI to preserve healthcare privacy. (...)