Update Detection Class for TensorFlow 2.x and NumPy Compatibility

[i-ninte]

Pull Request Description:
This update refactors the Detection class to ensure compatibility with TensorFlow 2.x and current NumPy standards. Key changes include:

• Replaced deprecated np.int and np.float types with standard int and float, aligning with modern NumPy practices.
• Ensured confidence is explicitly set as float and feature uses np.float32 for optimal memory usage and compatibility with neural network models.
• Refactored code for improved readability and adherence to Python 3 object-oriented practices.
  -Updated the import statement to utilize linear_sum_assignment from scipy.optimize, ensuring efficient assignment in optimization tasks