Critiscan's Fruit Quality Assessment Model is an advanced deep learning solution for detecting the freshness of fruits and vegetables using computer vision. Developed for the Flipkart Robotics Challenge Hackathon, this model provides an innovative approach to quality assessment in agricultural produce.

• Link to Website Repository
• Link to OCR Detection Model
• Link to Item Counting and Brand Detection

• High-Accuracy Freshness Classification: Achieves an impressive 99.69% accuracy in detecting fruit and vegetable quality
• Multi-Class Detection: Classifies produce across six fruit types with granular quality states
• Robust Deep Learning Architecture: Utilizes state-of-the-art EfficientNetB5 model with advanced regularization techniques

• Source: FruitNet Indian Fruits Dataset with Quality
• Fruits Covered: Apple, Banana, Guava, Lime, Orange, Pomegranate
• Quality States: Good, Bad, Mixed

• Base Model: EfficientNetB5 (pre-trained on ImageNet).
• Input Shape: (img_size[0], img_size[1], 3)
• Modifications:
  • Batch normalization layers added for regularization.
  • Two dense layers (1024 and 128 neurons) with L1/L2 regularization.
  • Dropout layers with rates of 30% and 45% to prevent overfitting.
  • Final output layer with softmax activation for multi-class classification.
• Optimizer: Adamax with an initial learning rate of 0.001.
• Loss Function: Categorical Crossentropy.

• Epochs: 40
• Batch Size: 20
• Data Augmentation: (Mention if used and details, e.g., rotation, zoom, flip.)
• Callbacks:
  • ReduceLROnPlateau: Reduces learning rate by a factor of 0.5 if validation loss stagnates for 2 epochs.
  • EarlyStopping: Stops training after 4 epochs of no improvement in validation loss.
  • ASK Callback: Interactive callback to query user on training continuation after every ask_epoch (5 epochs in this case).

• Framework: Flask
• Containerization: Docker
• Cloud Deployment: AWS EC2
• Inference Time: 3-4 seconds per prediction

• Python 3.8+
• pip
• Docker (optional)

#Clone the repository:    
    git clone https://github.com/tsu-ki/Freshness-model
    
#Install dependencies:
    pip install -r requirements.txt
    
#Run the application:
   python app.py

• Expand dataset diversity
• Include packaged food freshness detection
• Optimize real-time inference capabilities

• Dataset: FruitNet: Indian Fruits Dataset with Quality
• EfficientNet Research Paper