CS6018 Image Processing

Case Studies in
Image Security

Presented by J B Mugundh
Introduction to
Image Security
Defi nition: Securing images against
unauthorized access, manipulation, and
misuse.

Importance: Protects sensitive data, privacy,

intellectual property, and image integrity.

Fields of Application: Social media, medical

imaging, digital forensics, government
databases.
 01. Confidentiality: Ensuring only
authorized users access images.

02.
Integrity: Preventing

Objectives of unauthorized
images.
alterations to

Image
03.
Authentication: Verifying the
origin and authenticity of

02
Security images.

04.
Non-repudiation: Ensuring
actions cannot be denied by
parties involved.
Encryptio
n
Transforming
into
formats
RSA).
images
unreadable
(e.g., AES,
Steganograp
hy
Concealing
within
covert
images

communication.
data
for
03
Watermarki
ng
Hashing Image
Embedding
information
hidden
for
Generating
identifiers to
unique
check
Security
Techniques
copyright protection. image integrity.
 Facial Recognition
Security in Financial
Services
Challeng Solution Result
e s
• Secure Access Control for • MFA: Combining facial • Security Enhancement:
banking systems and ATMs. recognition with PINs or Reduction in unauthorized
mobile authentication. access attempts by 70%.
• Preventing unauthorized
Case Study -

access through spoofi ng • Liveness Detection: • User Convenience: Faster

and presentation attacks.
Utilizing eye-blink and contactless
detection, 3D depth authentication process
• Ensuring user privacy and
compliance with data sensing, and texture improved customer
protection regulations. analysis to diff erentiate satisfaction.
between real users and
spoofed images. • Regulatory Compliance:
Achieved compliance with
• Secure Biometric GDPR and other data
Storage: Encrypting facial protection laws.
data and storing it in
secure databases to
prevent breaches.
1
 Medical Imaging Security
in Healthcare
Challeng Solutio Results
e n
• Protecting patient privacy • Encryption of Medical • Improved Patient
and sensitive medical Images: Utilizing AES-256 Privacy: Enhanced
images in healthcare encryption to protect protection of sensitive
systems. images both at rest and in medical data, leading to
Case Study -

transit. increased patient trust.

• Safeguarding medical
images (e.g., X-rays, MRIs)
• Access Controls: • Regulatory Compliance:
from unauthorized access
and breaches. Implementing role-based Successfully met HIPAA
access control (RBAC) to requirements, avoiding
• Ensuring compliance with ensure only authorized potential fi nes and legal
healthcare regulations like medical personnel can issues.
HIPAA. view or modify images.
• Data Integrity: Maintained
• Audit Trails: Maintaining the accuracy and reliability
detailed logs of who of medical images, crucial
accessed or altered for patient diagnosis and
medical images to ensure treatment.
2

accountability.
 Blockchain for Image
Integrity in Supply Chain
Management
Challeng Solutio Results
e n
• Ensuring the authenticity • Blockchain Integration: • Enhanced Trust: Increased
and traceability of product Recording hashes of confi dence among
images in the supply chain. images on a blockchain consumers and partners
ledger to create an regarding the authenticity
Case Study -

• Preventing image immutable record of their of product images.

tampering and verifying authenticity and history.
the origin of product
• Tamper-Proof Records:
images used in marketing
• Smart Contracts: Prevented unauthorized
and documentation.
Automating verifi cation alterations and ensured a
• Maintaining a transparent processes and access reliable audit trail for image
and immutable record of controls for image usage usage.
image usage and based on predefi ned rules.
modifi cations. • Operational Effi ciency:
• Decentralized Storage: Streamlined verifi cation
Utilizing decentralized processes, reducing the
storage solutions like IPFS time and resources needed
to store images securely for image authentication.
3

and redundantly.
 Threats Security

0 01
GAN-Based
Deepfake
Watermarking: Using
Manipulation:
Generative Adversarial
Misusing AI to create
Networks to create hard-

1
realistic fake images.
to-detect watermarks.

02 Technolo
02
Blockchain for Image
Image-based
Integrity:
Phishing

gy
Decentralized tracking
(Spearfishing): Advancements of image history and
Creating fake images
usage.
for fraud.

03
07
03
Quantum Encryption:
AI-Driven Attacks: Using QKD for virtually
Using AI to bypass unbreakable
traditional security encryption, protecting
measures. images and data
Conclusio
n
• Image security is critical for safeguarding
privacy, ensuring data integrity, and
protecting intellectual property across
various industries.

• Ongoing advancements in AI, blockchain,

cryptography, and quantum computing
will continue to shape and enhance the
landscape of image security.

• Investing in comprehensive image

security measures is essential for
protecting digital assets, maintaining
trust, and enabling secure digital
interactions in an increasingly image-
driven world.
Thank
You
Appendix
Steganography Encoding
1. from PIL import Image 16.
2. def encode_image(image_path, message, 17. # Modify the least signifi cant bit of
output_path): each color component
3. # Open the input image 18. r = (r & ~1) |
4. img = Image.open(image_path) int(binary_message[index]) # Red channel
5. img = img.convert('RGB') # Ensure image is LSB
in RGB mode 19. if index + 1 < len(binary_message):
6. pixels = img.load() 20. g = (g & ~1) |
int(binary_message[index + 1]) # Green
7. # Convert the message to binary channel LSB
8. binary_message = ''.join([format(ord(char), 21. if index + 2 < len(binary_message):
'08b') for char in message]) + '00000000' # 22. b = (b & ~1) |
End with 8 zeros as a terminator int(binary_message[index + 2]) # Blue channel
9. LSB
10. # Encoding the message into pixels 23.
11. index = 0 24. pixels[x, y] = (r, g, b)
12. for y in range(img.height): 25. index += 3
13. for x in range(img.width): 26. else:
14. if index < len(binary_message): 27. break
15. r, g, b = pixels[x, y]
28. # Save the encoded image
Steganography Decoding
1. def decode_image(image_path): 14. # Convert the binary message to text
2. # Open the encoded image 15. message = ''
3. img = Image.open(image_path) 16. for i in range(0, len(binary_message), 8):
4. img = img.convert('RGB') 17. byte = binary_message[i:i + 8]
5. pixels = img.load() 18. if byte == '00000000': # End of message
terminator
6. # Extracting the binary message from the image 19. break
7. binary_message = '' 20. message += chr(int(byte, 2))
8. for y in range(img.height):
9. for x in range(img.width): 21. print(f"Decoded message: {message}")
10. r, g, b = pixels[x, y] 22. return message
11. binary_message += str(r & 1) # Red
channel LSB
12. binary_message += str(g & 1) # Green
channel LSB
13. binary_message += str(b & 1) # Blue
channel LSB