I Built a CNN to Detect Skin Cancer from Images (Beginner ML Project)

Hey everyone 👋

Just wanted to share a machine learning project I recently built as part of my learning journey. It's a basic skin cancer detection model using a Convolutional Neural Network (CNN). The model classifies skin lesion images as benign or malignant, and I tested it locally with a Streamlit app.

Why I Picked This Project

I’m still learning machine learning, and I wanted to try something practical something where I could take an idea, build a model, and test it with real images. Skin cancer is a serious health issue, and early detection helps a lot, so I thought this would be a good starting point for a classification task.

⚠️ Disclaimer: This is an educational project only not for real medical use.

Tools I Used

Python
TensorFlow + Keras
Streamlit
Pillow / NumPy
Jupyter Notebook

What the Model Does

Loads a skin lesion image
Preprocesses it (resize, normalize)
Predicts if the image is benign or malignant
Shows the result in a local Streamlit interface

CNN Architecture (Simplified)

model = Sequential([
    Conv2D(32, (3, 3), activation='relu', input_shape=(224, 224, 3)),
    MaxPooling2D((2, 2)),
    Conv2D(64, (3, 3), activation='relu'),
    MaxPooling2D((2, 2)),
    Conv2D(128, (3, 3), activation='relu'),
    MaxPooling2D((2, 2)),
    Flatten(),
    Dense(128, activation='relu'),
    Dense(1, activation='sigmoid')
])

Trained with binary cross-entropy since it's a binary classification task.

Prediction Function (Streamlit)

def predict_skin_cancer(image_path, model):
    img = image.load_img(image_path, target_size=(224, 224))
    img_array = image.img_to_array(img) / 255.0
    img_array = np.expand_dims(img_array, axis=0)

    prediction = model.predict(img_array)
    return "Malignant" if prediction > 0.5 else "Benign"