CalvinYu727/invoice_cheque_creditcard_model

Model Card for Model ID

A fine-tuned model for classifying invoices, credit card and bank cheque, trained using MobileNetV3_Large using Google Colab.

Model Description

• Developed by: Calvin
• Model type: Image Classification
• License: MIT
• Finetuned from model: timm/mobilenetv3_large_100.ra_in1k

''' from tensorflow.keras.models import load_model import cv2 from tensorflow.keras.preprocessing.image import img_to_array import tensorflow as tf import numpy as np from google.colab import files, drive import matplotlib.pyplot as plt

Define image size

IMG_SIZE = 224

Set confidence threshold

CONFIDENCE_THRESHOLD = 0.5

drive.mount('/content/drive')

Load the finetuned model

finetuned_model = load_model('/content/drive/MyDrive/models/finetuned_mobilenetv3large_model_v4.keras')

Upload image using google.colab.files

uploaded = files.upload()

Process the uploaded image

for fn in uploaded.keys(): # Read the image img_path = fn img = cv2.imread(img_path)

  # Resize the image (using tf.image.resize)
  img = tf.image.resize(img, [IMG_SIZE, IMG_SIZE])

  # Convert image to array
  img_array = img_to_array(img)

  # Normalize pixel values
  img_array = img_array / 255.0

  # Add a dimension (batch size)
  img_array = tf.expand_dims(img_array, 0)

  # Perform inference using the finetuned model
  prediction = finetuned_model.predict(img_array)

  # Get probabilities for all classes
  class_probabilities = prediction[0]

  # Display results
  print(f'{fn}:')

  max_prob = np.max(class_probabilities)
  if max_prob > CONFIDENCE_THRESHOLD:
      for i, prob in enumerate(class_probabilities):
          if i == 0:
              class_name = 'Bank Cheque'
          elif i == 1:
              class_name = 'Credit Card'
          elif i == 2:
              class_name = 'Invoice'
          else:
              class_name = 'Others'
          print(f' {class_name}: ({prob:.4f})')
  else:
      print('Others')

  # Display the image
  plt.imshow(img.numpy().astype(np.uint8))
  plt.title(fn)
  plt.show()