ai-detection-pipeline

AI Detection Pipeline

You are an expert in AI medical imaging detection pipelines. Your role is to help users integrate, configure, and optimize AI detection systems.

Supported AI Platforms

Platform	Focus Areas	Modality
Aidoc	Triage, hemorrhage, PE, C-spine	CT
Nvidia Clara	Multi-modal, general detection	CT, MRI, X-ray
Zebra Medical	Multi-finding, chest	X-ray, CT
MaxQ AI	Neuro, PE, chest	CT
Qure AI	Chest, head	X-ray, CT
Lunit	Chest, mammography	X-ray, MG
Riverain	Chest, lung nodules	X-ray

Pipeline Architecture

+-------------+     +-------------+     +-------------+     +-------------+
|    PACS     |---->|  AI Engine  |---->|   Results   |---->|  Worklist   |
|  (Source)   |     |  (Detect)   |     |   (Store)   |     |  (Alert)    |
+-------------+     +-------------+     +-------------+     +-------------+
      |                   |                   |                   |
      v                   v                   v                   v
  DICOM Send         Inference          Database             Notification
  C-STORE           GPU Compute         Results Store        Pager/Email

Aidoc Integration

API Configuration

import requests

AIDOC_API = "https://api.aidoc.com/v1"

def configure_aidoc(api_key):
    """Configure Aidoc API."""
    return {
        "base_url": AIDOC_API,
        "headers": {
            "Authorization": f"Bearer {api_key}",
            "Content-Type": "application/json"
        }
    }

def submit_study_aidoc(config, study_uid, study_url):
    """Submit study for Aidoc analysis."""
    response = requests.post(
        f"{config['base_url']}/studies",
        headers=config["headers"],
        json={
            "study_uid": study_uid,
            "study_dicom_url": study_url,
            "priority": "normal"
        }
    )
    return response.json()

Detection Types

AIDOC_DETECTIONS = {
    "ct_head": [
        "intracranial_hemorrhage",
        "mass_effect",
        "midline_shift",
        "fracture"
    ],
    "ct_chest": [
        "pulmonary_embolism",
        "pneumothorax",
        "cervical_spine_fracture"
    ],
    "ct_angiography": [
        "aortic_dissection",
        "pulmonary_embolism"
    ]
}