SR conversion to encapsulated PDF document DICOM object

SRToPDF.py

@@ -0,0 +1,148 @@
+import pydicom
+from pydicom.dataset import Dataset
+from pydicom.dataset import FileMetaDataset
+from pydicom.uid import MediaStorageDirectoryStorage, EncapsulatedPDFStorage, generate_uid
+import matplotlib.pyplot as plt
+from reportlab.pdfgen import canvas
+
+
+def load_mammography_image(dicom_path):
+    """Loads a mammography DICOM image as a NumPy array."""
+    ds = pydicom.dcmread(dicom_path)
+    image = ds.pixel_array
+    return image, ds
+
+
+def extract_measurements(sr_path):
+    """Extracts measurement annotations from an SR DICOM file."""
+    ds = pydicom.dcmread(sr_path)
+    measurements = []
+    probabilities = []
+
+    if "ContentSequence" in ds:
+        for itemLevel1 in ds.ContentSequence:
+            if len(itemLevel1.ConceptNameCodeSequence) ==  1:
+                if itemLevel1.ConceptNameCodeSequence[0].CodeMeaning == "Imaging Measurements":
+                    for itemLevel2 in itemLevel1.ContentSequence:
+                        for itemLevel3 in itemLevel2.ContentSequence:
+                            if itemLevel3.ValueType == "SCOORD":
+                                measurements.append(itemLevel3.GraphicData)
+                            elif itemLevel3.ValueType == "NUM":
+                                if len(itemLevel3.MeasuredValueSequence) == 1:
+                                    probabilities.append(itemLevel3.MeasuredValueSequence[0].NumericValue)
+    return measurements, probabilities
+
+
+def overlay_measurements(image, measurements, probabilities):
+    """Overlays extracted measurements onto the mammography image."""
+    fig, ax = plt.subplots()
+    ax.imshow(image, cmap='gray')
+
+    # Draw each polyline
+    for i in range(0, len(measurements), 1):
+        measurement = measurements[i]
+        x = measurement[0::2]  # Extract x-coordinates (every other value)
+        y = measurement[1::2]  # Extract y-coordinates (every other value)
+        ax.plot(x, y, 'lime', linewidth=1)  # Plot the entire polyline at once
+        ax.text(x[-3] + 100, y[-3], f"{probabilities[i]:.2f} %", color='lime', fontsize=8)
+
+    ax.axis("off")
+
+    # Save the overlay as an image
+    plt.savefig("temp.png", bbox_inches='tight', pad_inches=0)
+    plt.close()
+
+
+def create_pdf(image_path, measurements, pdf_path):
+    """Creates a PDF with the mammography image and extracted measurements."""
+    c = canvas.Canvas(pdf_path)
+
+    # Add the image to the PDF
+    c.drawImage(image_path, 50, 300)
+
+    # Add extracted measurements
+    c.setFont("Helvetica", 12)
+    c.drawString(50, 280, "Extracted Measurements:")
+
+    y = 260
+    for text in measurements:
+        c.drawString(70, y, f"- {text}")
+        y -= 15  # Move down for each line
+
+    c.save()
+
+def create_dcm_pdf(sr_path, pdf_path):
+    sr = pydicom.dcmread(sr_path)
+    ds = Dataset()
+
+    # Add general DICOM metadata
+    ds.PatientName = sr.PatientName
+    ds.PatientID = sr.PatientID
+    ds.PatientBirthDate = sr.PatientBirthDate
+    ds.PatientSex = sr.PatientSex
+
+    ds.StudyInstanceUID = sr.StudyInstanceUID
+    ds.StudyDate = sr.StudyDate
+    ds.StudyTime = sr.StudyTime
+    ds.AccessionNumber = sr.AccessionNumber
+    ds.ReferringPhysicianName = sr.ReferringPhysicianName
+    ds.StudyID = sr.StudyID
+
+    ds.SeriesInstanceUID = generate_uid()
+    ds.SeriesDate = sr.SeriesDate
+    ds.SeriesTime = sr.SeriesTime
+    ds.SeriesNumber = 1
+    ds.Modality = "DOC"
+
+    ds.Manufacturer = "MammographyAI"
+    ds.ConversionType = "DI"
+
+    ds.SOPInstanceUID = generate_uid()
+    ds.SOPClassUID = EncapsulatedPDFStorage
+
+    # Open the PDF file and read it as binary data
+    with open(pdf_path, 'rb') as f:
+        pdf_data = f.read()
+
+    # Add the EncapsulatedDocument (PDF content) to the DICOM dataset
+    ds.ContentDate = ds.SeriesDate
+    ds.ContentTime = ds.SeriesTime
+    ds.AcquisitionDateTime = ""
+    ds.InstanceNumber = 1
+    ds.BurnedInAnnotation = "YES"
+    ds.DocumentTitle = ""
+    ds.EncapsulatedDocument = pdf_data
+    ds.MIMETypeOfEncapsulatedDocument = "application/pdf"
+
+    # Create a FileMetaDataset for DICOM file meta information
+    file_meta = FileMetaDataset()
+    file_meta.MediaStorageSOPClassUID = EncapsulatedPDFStorage
+    file_meta.MediaStorageSOPInstanceUID = ds.SOPInstanceUID
+    file_meta.TransferSyntaxUID = pydicom.uid.ImplicitVRLittleEndian
+    file_meta.FileMetaInformationGroupLength = 0
+
+    # Assign the file meta information to the dataset
+    ds.file_meta = file_meta
+
+    # Ensure preamble and "DICM" prefix is included
+    ds.is_implicit_VR = True  # Set to explicit VR
+    ds.is_little_endian = True  # Set to little endian
+
+    # Save the DICOM file with the preamble and DICM prefix
+    output_file = "output.dcm"
+    ds.save_as(output_file, write_like_original=False)
+
+    print(f"DICOM file created: {output_file}")
+
+    test = pydicom.dcmread("output.dcm")
+    print("test")
+
+# Example usage
+mammogram_path = "MG000001.dcm"
+sr_path = "SR000001.dcm"
+
+image, dicom_ds = load_mammography_image(mammogram_path)
+measurements, probabilities = extract_measurements(sr_path)
+overlay_measurements(image, measurements, probabilities)
+create_pdf("temp.png", measurements, "temp.pdf")
+create_dcm_pdf(sr_path, "temp.pdf",)