Handwritten digit recognition based on ghost imaging and cake-cutting Hadamard matrix

Key Points

• The method achieves high accuracy in recognizing handwritten digits from ghost images, supporting its effectiveness.
• Key results indicate over 95% recognition accuracy in experimental trials with various digit datasets.
• The approach leverages ghost imaging combined with a Hadamard matrix to enhance image processing in machine learning.
• This technique may improve capabilities in optical imaging systems, highlighting potential for future applications.