medthermal-dicom 1.0.4

Thermal imaging DICOM library for medical applications

MedThermal DICOM

Thermal imaging DICOM library for medical applications

A comprehensive Python library for creating and managing thermal DICOM images with support for thermal-specific metadata.

Table of Contents

• Overview
• Features
• Installation
• Quick Start
• Converting FLIR Radiometric Images to CSV
• Usage Examples
  • Basic DICOM Creation
  • Setting Thermal Parameters
  • Metadata Management
  • Adding Overlays for ROI Visualization
• GUI Application
• Examples
• License

Overview

MedThermal DICOM is designed for researchers, clinicians, and developers working with medical thermal imaging. It provides:

• Python API: Programmatic creation and manipulation of thermal DICOM files
• GUI Applications: User-friendly interfaces for non-programmers
• Standards Compliance: Full DICOM compliance with thermal-specific extensions
• Rich Metadata: Comprehensive thermal imaging metadata support

Features

Core Features

• ✅ Create DICOM files from thermal images
• ✅ Import temperature data from CSV or numpy arrays or JPEG or PNG
• ✅ Set key thermal acquistion parameters (emissivity, distance, ambient temperature, etc.)
• ✅ Manage patient, study, and series metadata
• ✅ Add binary mask overlays for ROI visualization
• ✅ Export fully compliant thermal images to DICOM
• ✅ Organization UID management for custom UID generation

GUI Features

• 🎨 Simple user interface to quickly convert Thermal Image to DICOM
• 🏥 Comprehensive patient and study information entry
• 🌡️ Thermal parameter configuration
• 📊 Organization UID management

Installation

Prerequisites

• Python 3.8 or higher
• pip package manager

Install as Package from PyPI

pip install medthermal-dicom

Install from Source

1. Clone or download this repository:

git clone https://github.com/medthermal/MedThermalDicom.git
cd MedThermalDicom

2. Install dependencies:

pip install -r requirements.txt

pip install -e .

This installs the medthermal_dicom package and makes it available system-wide.

Quick Start

Using the API

Here's a minimal example to create a DICOM file with metadata and body part:

import numpy as np
from medthermal_dicom import MedThermalDicom, MedThermalMetadata

# Create thermal DICOM
thermal_dicom = MedThermalDicom()
temperature_data = np.random.rand(256, 256) * 10 + 30  # Sample temperature data
thermal_dicom.set_thermal_image(temperature_data,  (30.0, 40.0))

# Add metadata and set body part
metadata = MedThermalMetadata()
metadata.set_patient_information(patient_name="TEST^PATIENT", patient_id="TH001")
metadata.set_study_information(study_description="Thermal Study")
metadata.set_series_information(series_description="Chest Imaging", body_part="chest")
metadata.apply_metadata_to_dataset(thermal_dicom.dataset)

# Save DICOM file
thermal_dicom.save_dicom("output.dcm")

Converting FLIR Radiometric Images to CSV

If you have radiometric JPG images from FLIR thermal cameras, you need to convert them to CSV format before using them with MedThermal DICOM.

Step 1: Download FLIR Tools

• Visit FLIR Tools Download Page
• Download and install FLIR Tools

Step 2: Open Radiometric JPG in FLIR Tools

• Launch FLIR Tools
• Open your radiometric JPG file

Step 3: Export to CSV

• Right-click on the thermal image
• Select "Export to CSV"
• Save the CSV file to your desired location
• Use CSV with MedThermal DICOM as shown in the Examples or GUI

Usage Examples

Basic DICOM Creation

from medthermal_dicom import MedThermalDicom
import numpy as np

# Create instance
thermal_dicom = MedThermalDicom()

# Load temperature data from CSV
temperature_data = np.loadtxt("temp_data.csv", delimiter=",")

# Set thermal image (display array, temperature array, temperature range)
temp_min, temp_max = temperature_data.min(), temperature_data.max()
thermal_dicom.set_thermal_image(
    thermal_array=temperature_data,
    temperature_range=(temp_min, temp_max)
)

Setting Thermal Parameters

# Assuming thermal_dicom is already created (see Basic DICOM Creation)

# Define thermal parameters
thermal_params = {
    'emissivity': 0.98,                    # Human skin emissivity
    'distance_from_camera': 1.0,           # Distance in meters
    'ambient_temperature': 22.0,           # Room temperature in °C
    'reflected_temperature': 22.0,         # Reflected temperature
    'atmospheric_temperature': 22.0,       # Atmospheric temp
    'relative_humidity': 45.0,            # Humidity percentage
    'temperature_range_min': 20.0,        # Min temperature
    'temperature_range_max': 40.0,        # Max temperature
    'temperature_unit': 'Celsius',        # Temperature unit
    'thermal_sensitivity': 0.05,          # NETD in °C
    'spectral_range': '7.5-14.0',         # Spectral range in μm
    'lens_field_of_view': 24.0            # FOV in degrees
}

thermal_dicom.set_thermal_parameters(thermal_params)

Metadata Management

from medthermal_dicom import MedThermalMetadata, MedThermalDicom
import numpy as np

# Create thermal_dicom instance
thermal_dicom = MedThermalDicom()
temperature_data = np.random.rand(256, 256) * 20 + 25
thermal_dicom.set_thermal_image(temperature_data,  
                               (temperature_data.min(), temperature_data.max()))

# Create metadata handler
metadata = MedThermalMetadata()

# Set patient information
metadata.set_patient_information(
    patient_name="ANONYMOUS^PATIENT",
    patient_id="TH001",
    patient_birth_date="19850315",
    patient_sex="M",
    patient_age="038Y"
)

# Set study information
metadata.set_study_information(
    study_description="Breast Thermal Imaging Study",
    accession_number="ACC123456",
    referring_physician="DR^EXAMPLE^PHYSICIAN",
    procedure_code="breast_thermography"  # SNOMED CT code
)

# Set series information
metadata.set_series_information(
    series_description="Thermal Images - Anterior View",
    body_part="breast",  # Uses SNOMED CT codes
    patient_position="HFS"
)

# Set equipment information
metadata.set_equipment_information(
    manufacturer="FLIR Systems",
    manufacturer_model="T1K",
    device_serial_number="SN12345",
    software_version="MedThermalDICOM v1.0"
)

# Apply metadata to DICOM dataset
metadata.apply_metadata_to_dataset(thermal_dicom.dataset)

# Save DICOM file
thermal_dicom.save_dicom("output.dcm")

Adding Overlays for ROI Visualization

You can add binary mask overlays to highlight regions of interest (ROI) in thermal images:

import numpy as np
from medthermal_dicom import MedThermalDicom

# Load thermal data
temperature_data = np.loadtxt("thermal_data.csv", delimiter=",")

# Create thermal DICOM
thermal_dicom = MedThermalDicom()
thermal_dicom.set_thermal_image(temperature_data)

# Create binary mask for ROI (same shape as thermal data)
roi_mask = np.zeros_like(temperature_data, dtype=bool)
roi_mask[100:200, 150:250] = True  # Define rectangular ROI

# Add overlay to DICOM
thermal_dicom.add_overlay(roi_mask)

# Set patient info and thermal parameters
thermal_dicom.create_standard_thermal_dicom(
    patient_name="TEST^PATIENT",
    patient_id="THERMAL001",
    study_description="Thermal Imaging with ROI",
    thermal_params={
        'emissivity': 0.98,
        'distance_from_camera': 1.0,
        'ambient_temperature': 22.5
    }
)

# Save DICOM file with overlay
thermal_dicom.save_dicom("output_with_overlay.dcm")

Key Points:

• The mask must be a boolean numpy array with the same shape as the thermal image
• Overlay pixels appear highlighted in DICOM viewers that support overlays
• Overlays are stored in DICOM-compliant overlay planes

GUI Application

The simple GUI provides an intuitive interface for creating single thermal DICOM files.

Windows:

Unzip methermaldicom_gui.zip file

methermaldicom_gui.exe

Launch:

python simple_dicom_gui.py

Features:

• Single file processing
• Patient information entry
• Thermal parameter configuration
• Organization UID selection

Workflow:

1. Browse and select input file (image or CSV)
2. Fill in patient information (Name, ID, Age, Gender)
3. Configure thermal parameters (optional)
4. Select output folder
5. Click "Create DICOM"

Examples

The examples/ directory contains comprehensive usage examples:

• basic_usage.py: Complete API tutorial covering all features
• medical_thermal_imaging.py: Medical imaging workflow example
• organization_uid_example.py: Organization UID management
• pixel_data_example.py: Advanced pixel data handling
• overlay_example.py: Add binary mask overlays for ROI visualization
• multi_view_thermal_imaging.py: Multi-view thermal imaging with different anatomical views

GUI Dependencies

tkinter                 # GUI framework (usually included with Python)

Installation

Install all dependencies:

pip install -r requirements.txt

For GUI only:

pip install -r gui_requirements.txt

License

This project is licensed under the MIT License. See LICENSE file for details.

Support

For questions, issues, or feature requests:

• Check existing documentation in the examples/ directory
• Review additional guides: ORGANIZATION_UID_GUIDE.md, GUI_README.md
• Open an issue on the project repository

Citation

If you use this code in your research, please cite our work using the following reference:

Govindaraju, B., Kakileti, S.T., Dedhiya, R., Manjunath, G. (2026). MedThermal-DICOM: An Open-Source DICOM-Compliant Framework for Medical Thermal Imaging Enabling Clinical Integration and Research Reproducibility. In: Kakileti, S.T., Manjunath, G., Schwartz, R.G., Ng, E.Y. (eds) Artificial Intelligence over Infrared Images for Medical Applications. AIIIMA 2025. Lecture Notes in Computer Science, vol 16308. Springer, Cham. https://doi.org/10.1007/978-3-032-10990-3_13

Acknowledgments

We gratefully acknowledge:

• DICOM Standards Committee - For establishing and maintaining comprehensive medical imaging standards that enable interoperability across healthcare systems
• PyDICOM Project - For providing excellent Python tools for working with DICOM files, which form the foundation of this library
• Medical Thermal Imaging Community - For advancing research and clinical applications in thermal imaging

Contributing

We are volunteers developing MedThermal-DICOM in our free time, and with limited resources we greatly appreciate any contributions—whether bug fixes, new features, or documentation enhancements.
For any queries or contributions, please write to [e-mail]