2024_Ruthe_SND/code/overlayPlot3D.py

import nibabel as nib
import numpy as np
import matplotlib.pyplot as plt
from scipy.ndimage import rotate

# Define the file paths
templateFilePath = r"C:\Users\aswen\Desktop\Code\2024_Ruthe_SND\input\average_template_50_from_website.nii.gz"
densityFilePath = r"C:\Users\aswen\Desktop\Code\2024_Ruthe_SND\input\12_wks_coronal_100141780_50um_projection_density.nii.gz"
maskFilePath = r"C:\Users\aswen\Desktop\Code\2024_Ruthe_SND\output\Nifti_Trakte_registered\CC_Mop_dilated_registered.nii.gz"

# Load the NIfTI files
templateNifti = nib.load(templateFilePath)
densityNifti = nib.load(densityFilePath)
maskNifti = nib.load(maskFilePath)

# Ensure that both images have the same data type (e.g., cast density to the data type of template)
densityNifti = nib.Nifti1Image(densityNifti.get_fdata().astype(templateNifti.get_fdata().dtype), densityNifti.affine)

# Customize the figure DPI
dpi = 400

# Get the number of slices along each dimension
num_slices_dim0, num_slices_dim1, num_slices_dim2 = templateNifti.shape

# Create a 3D array to hold the overlay
overlay_volume = np.zeros((num_slices_dim0, num_slices_dim1, num_slices_dim2, 4), dtype=np.uint8)

# Loop through and overlay all slices
for sliceIdx in range(num_slices_dim2):
    # Extract the desired slice from the template, density, and mask
    templateSlice = np.squeeze(templateNifti.get_fdata()[:, :, sliceIdx]).astype(np.uint16)
    densitySlice = np.squeeze(densityNifti.get_fdata()[:, :, sliceIdx]).astype(np.double)
    
    mask_data = maskNifti.get_fdata()
    mask_data = mask_data.swapaxes(0, 2)
    mask_data = np.flip(mask_data, axis=0)

    maskSlice = np.squeeze(mask_data[:, :, sliceIdx])

    # Normalize the templateSlice to [0, 1] for visualization
    templateSlice = (templateSlice - np.min(templateSlice)) / (np.max(templateSlice) - np.min(templateSlice))

    # Overlay the density and mask onto the template slice
    overlayedImage = templateSlice + densitySlice + maskSlice

    # Normalize the overlay to [0, 1]
    overlayedImage = (overlayedImage - np.min(overlayedImage)) / (np.max(overlayedImage) - np.min(overlayedImage))

    # Convert the overlay image to RGBA format (4 channels for transparency)
    overlay_image_rgba = plt.cm.gray(overlayedImage)
    overlay_image_rgba[:, :, 3] = (overlayedImage * 255).astype(np.uint8)  # Set alpha channel based on intensity

    # Update the 3D overlay volume
    overlay_volume[:, :, sliceIdx, :] = overlay_image_rgba

# Create a figure for the 3D visualization
fig = plt.figure(figsize=(10, 10), dpi=dpi)
ax = fig.add_subplot(111, projection='3d')

# Plot the 3D overlay volume
ax.voxels(overlay_volume[:, :, :, 0] > 0, facecolors=overlay_volume[:, :, :, :3] / 255., edgecolor='k')

# Set the aspect ratio to be equal for all dimensions
ax.set_box_aspect([1, 1, 1])

# Show the 3D plot
plt.show()