Breast Histopathology Image Segmentation Part 6
- Part 1: Data Inspection and Pre-processing
- Part 2: Weights, Data Augmentations and Generators
- Part 3: Model creation based on a pre-trained and a custom model
- Part 4: Train our model to fit the dataset
- Part 5: Evaluate the performance of your trained model
- Part 6: Running Predictions
Based on Breast Histopathology Images by Paul Mooney.
Invasive Ductal Carcinoma (IDC) is the most common subtype of all breast cancers. To assign an aggressiveness grade to a whole mount sample, pathologists typically focus on the regions which contain the IDC. As a result, one of the common pre-processing steps for automatic aggressiveness grading is to delineate the exact regions of IDC inside of a whole mount slide.Can recurring breast cancer be spotted with AI tech? - BBC News
- Citation: Deep learning for digital pathology image analysis: A comprehensive tutorial with selected use cases
- Dataset: 198,738 IDC(negative) image patches; 78,786 IDC(positive) image patches
Loading the Model
After fitting the model to classify our source images into malignant and benign we can now use this model to make predictions about un-classified images. So load the model we want to use:
./RunPrediction.py
# Model you want to use
modelName = 'resnet50_weights.hdf5'
# modelName = 'custom_weights.hdf5'
modelPath = config.OUTPUT_PATH + '/' + modelName
print("Loading Breast Cancer detector model...")
model = load_model(modelPath)
Preparing the Test Image File
# Test image
# imagePath ="./sample_pictures/malignant.png"
imagePath = "./sample_pictures/benign.png"
# Loading the input image using openCV
image = cv2.imread(imagePath)
# Convert it from BGR to RGB and then resize it to 48x48,
# the same parameter we used for training
image1 = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
image1 = cv2.resize(image1, (48, 48))
image1 = img_to_array(image1)
# Only for ResNet50
image1 = preprocess_input(image1)
# The image is now represented by a NumPy array of shape (48, 48, 3), however
# we need the dimensions to be (1, 3, 48, 48) so we can pass it
# through the network and then we'll also preprocess the image by subtracting the
# mean RGB pixel intensity from the ImageNet dataset
image1 /= 255.0
image1 = np.expand_dims(image1, axis=0)
Make Predictions
# Pass the image through the model to determine if the person has malignant
(benign, malignant) = model.predict(image1)[0]
# Adding the probability in the label
label = "benign" if benign > malignant else "malignant"
label = "{}: {:.2f}%".format(label, max(benign, malignant) * 100)
# Showing the output image
print("RESULT :" +label)
cv2.imshow("IDC", image)
if cv2.waitKey(5000) & 0xFF == ord('q'):
cv2.destroyAllWindows()