Printer

- 2 mins read

IngeHack2k25 - reverse/mjsc Write-up

Challenge Overview

Challenge Name: reverse/printer
Category: Reverse Engineering
Event: IngeHack 2k25
Difficulty: Medium
Flag: ingehack{...}

Fourier Transformation-Based Image Decryption

Challenge Analysis

Upon analyzing the provided files, we observed the following directory structure:

➜  solver ls 
decrypted.png  enc  main  reconstructed_data.txt  sol.py

The main objective is to decrypt the file enc to retrieve an image. By inspecting main, we determined that it applies a Fourier transformation, meaning the encryption likely involves transforming the image into the frequency domain.

To recover the original image, we need to apply the Inverse Fourier Transform (IFFT) to enc and reconstruct the original pixel data.

Decryption Process

The decrypt_image() function in sol.py implements the decryption using NumPy’s FFT functions. Below is a step-by-step breakdown of how it works:

  1. Read the Encrypted File

    • The function reads enc as a binary file.
    • The binary data is interpreted as an array of float64 values.

  2. Reshape into Complex Numbers

    • The data is reshaped into pairs of real and imaginary components.
    • These are converted into complex numbers representing frequency-domain data.

  3. Apply Inverse FFT (IFFT)

    • The inverse Fourier transform is applied to reconstruct the time-domain (original pixel) data.

  4. Convert Back to Byte Values

    • The real part of the transformed data is extracted.
    • Values are rounded, clipped between 0-255, and converted to uint8 format.

  5. Save as PNG File

    • The decrypted image is written to decrypted.png.

Execution

To decrypt the image, simply run:

import numpy as np

def decrypt_image():
    # Read the encrypted file
    with open("enc", "rb") as f:
        data = f.read()
    
    # Convert to numpy array of doubles
    doubles = np.frombuffer(data, dtype=np.float64)
    
    # Reshape into complex numbers (real + imaginary pairs)
    complex_data = doubles.reshape(-1, 2)
    complex_numbers = complex_data[:, 0] + 1j * complex_data[:, 1]
    
    # Apply inverse FFT to get back to time domain
    time_domain_data = np.fft.ifft(complex_numbers)
    
    # Convert the real parts to bytes
    # Round and clip to ensure valid byte values (0-255)
    byte_data = np.clip(np.round(np.real(time_domain_data)), 0, 255).astype(np.uint8)
    
    # Write the decrypted data to a PNG file
    with open("decrypted.png", "wb") as f:
        f.write(bytes(byte_data))

if __name__ == "__main__":
    try:
        decrypt_image()
        print("Image decrypted successfully. Saved as 'decrypted.png'")
    except Exception as e:
        print(f"Error occurred: {e}")

Flag

After running the decryption, examining decrypted.png reveals the flag: decrypted