System Overview

PACS System is a modern C++20 PACS (Picture Archiving and Communication System) library for medical imaging. It implements the DICOM standard from scratch with zero external DICOM library dependency, providing complete SOP class support, transfer syntax handling, and DICOMweb services.

Built on the kcenon ecosystem infrastructure (networking, threading, logging, containers), PACS System delivers a layered architecture of 12 library targets covering the full DICOM pipeline from tag-level data handling through network services to cloud storage and AI integration.

Key Features

Category	Feature	Description
Core	DICOM Part 10 I/O	Native file read/write with full VR support
Core	12 Library Targets	Layered architecture from core to integration
Encoding	Transfer Syntaxes	Implicit/Explicit VR, JPEG, JPEG 2000, HTJ2K, JPEG-LS, RLE, HEVC
Network	DIMSE Services	C-ECHO, C-STORE, C-FIND, C-MOVE, C-GET, N-CREATE, N-SET
Network	Association Management	Full DICOM Upper Layer state machine
Web	DICOMweb	WADO-RS, STOW-RS, QIDO-RS via Crow REST framework
Storage	Multi-Backend	SQLite indexing, file storage, AWS S3, Azure Blob
AI	AI Integration	AI service connector with SR/SEG result handling
Monitoring	Prometheus Metrics	Health checks and Prometheus-compatible metrics export
Security	RBAC	Role-based access control and PS3.15 anonymization
Security	TLS	OpenSSL-based secure DICOM associations (BCP 195)
Workflow	Task Management	Auto prefetch, task scheduler, study lock

Architecture Diagram

The 12 libraries are organized in layers. Higher layers depend on lower layers, but never the reverse.

Quick Start

A minimal example reading a DICOM file and extracting patient information:

#include "kcenon/pacs/core/dicom_file.h"
#include "kcenon/pacs/core/dicom_dataset.h"
#include "kcenon/pacs/core/dicom_tag_constants.h"
 
#include <iostream>
 
int main() {
    // Read a DICOM Part 10 file
    auto result = kcenon::pacs::core::dicom_file::read("image.dcm");
    if (!result.is_ok()) {
        std::cerr << "Failed to read: " << result.error().message << "\n";
        return 1;
    }
 
    auto& file = result.value();
    auto& dataset = file.dataset();
 
    // Extract patient metadata
    auto patient_name = dataset.get_string(kcenon::pacs::core::tags::patient_name);
    auto study_date   = dataset.get_string(kcenon::pacs::core::tags::study_date);
 
    std::cout << "Patient: " << patient_name.value_or("(unknown)") << "\n";
    std::cout << "Study:   " << study_date.value_or("(unknown)") << "\n";
 
    return 0;
}

Installation

CMake FetchContent (Recommended)

include(FetchContent)
FetchContent_Declare(
    pacs_system
    GIT_REPOSITORY https://github.com/kcenon/pacs_system.git
    GIT_TAG        v0.1.0
)
FetchContent_MakeAvailable(pacs_system)
 
target_link_libraries(your_target PRIVATE
    kcenon::pacs_core
    kcenon::pacs_encoding
    kcenon::pacs_services
)

Note: PACS System requires DCMTK-compatible libraries and ICU for character set support. SQLite3 >= 3.45.1 is needed for storage features. Optional codec libraries (libjpeg-turbo, OpenJPEG, CharLS, OpenJPH) enable additional transfer syntax support.

vcpkg

vcpkg install kcenon-pacs-system

find_package(pacs_system CONFIG REQUIRED)

target_link_libraries(your_target PRIVATE kcenon::pacs_core)

Module Overview

Module	Directory	Description
pacs_core	`core/`	Tags, elements, datasets, Part 10 file I/O, DICOM dictionary
pacs_encoding	`encoding/`	VR types, transfer syntaxes, compression codecs (JPEG, JPEG 2000, HTJ2K, JPEG-LS, RLE)
pacs_security	`security/`	RBAC, PS3.15 anonymization, digital signatures, TLS profiles
pacs_network	`network/`	PDU encode/decode, association state machine, DIMSE pipeline
pacs_client	`client/`	Job management, routing, synchronization, prefetch, remote node management
pacs_services	`services/`	SCP/SCU implementations, SOP class registry, IOD validators
pacs_storage	`storage/`	File storage, SQLite indexing, cloud backends (AWS S3, Azure Blob), HSM
pacs_ai	`ai/`	AI service connector, SR/SEG result handler
pacs_monitoring	`monitoring/`	Health checks, Prometheus-compatible metrics export
pacs_workflow	`workflow/`	Auto prefetch, task scheduler, study lock
pacs_web	`web/`	REST API via Crow, DICOMweb (WADO-RS, STOW-RS, QIDO-RS)
pacs_integration	`integration/`	Adapters bridging PACS to the kcenon ecosystem

Examples

Example	Source	Description
dcm_extract	dcm_extract	Extract pixel data from DICOM files to JPEG, PNG, or raw binary
dcm_conv	dcm_conv	Convert DICOM files between transfer syntaxes
img_to_dcm	img_to_dcm	Encapsulate standard images (JPEG, PNG) into DICOM Part 10 files
query_scu	query_scu	Query a PACS server using C-FIND at Patient/Study/Series level
dcm_dump	dcm_dump	Dump DICOM file metadata as human-readable text
dcm_info	dcm_info	Display summary information for a DICOM file
echo_scu	echo_scu	Verify DICOM association with C-ECHO
store_scu	Step 2: Send a file from a C-STORE SCU (Storage User)	Send DICOM files to a PACS server via C-STORE
find_scu	find_scu	Query remote PACS with C-FIND
pacs_server	pacs_server	Full PACS server with SCP services

Learning Resources

New to pacs_system? Start here:

Related Systems

PACS System is the highest-level application (Tier 5) in the kcenon ecosystem. It depends on the following systems:

System	Tier	Role	Repository
common_system	0	Result<T>, IExecutor, patterns	https://github.com/kcenon/common_system
container_system	1	DICOM serialization support	https://github.com/kcenon/container_system
thread_system	1	Thread pools, async execution	https://github.com/kcenon/thread_system
logger_system	2	Structured logging	https://github.com/kcenon/logger_system
monitoring_system	2	Metrics collection	https://github.com/kcenon/monitoring_system
database_system	3	Database abstractions	https://github.com/kcenon/database_system
network_system	4	TCP/TLS, async networking	https://github.com/kcenon/network_system

Table of Contents