shahmi0519/fruit-ripeness-fusion-model

🍌 Fruit Ripeness Detection (Image + Tabular Early Fusion)

This model uses both images of bananas and simulated sensor readings (weight, moisture, days_since_harvest) to detect ripeness.

🧠 Architecture

• CNN (2 Conv layers + Flatten) → for image
• ANN (Dense) → for tabular
• Fusion → Concatenated features → Dense → Sigmoid (binary)

📦 Inputs

• 🖼️ 128×128 RGB image
• 📋 Tabular: weight (g), moisture (%), days since harvest

🎯 Output

• 0 = Fresh
• 1 = Rotten

🔧 Training

• Epochs: 15
• Loss: binary crossentropy
• Optimizer: Adam

🔍 Example Usage

from tensorflow.keras.models import load_model

model = load_model("fruit_ripeness_fusion_model.h5")