Search

Abstract: There has been remarkable growth in the digital domain. This has led to the production of a huge amount of digital data in the form of text messages, documents, audio/speech signals, digital images, videos, etc. In forensics, digital information (i.e., images of crime scenes, etc.) is very important to examine cases and collect the proof regarding the incident. This further leads to judge the victims and culprits by the court for appropriate actions. So, even a slight change (intentional/unintentional) may lead to the wrong judgement. Several digital signal processing tools are easily available to process this type of digital data, which means data may get disrupted, altered, or deform in such a way that the authenticity gets compromised. In such cases, it can be difficult to recognize the source information as well as to check the originality/intactness of the digital data. In the recent past, several techniques have come into existence to protect digital data against security issues. Steganography, cryptography, and watermarking are examples of security techniques, which have been widely used for data security and forensic applications for a long time. Specifically, digital watermarking techniques provide many useful features to help users regarding data security and authentication without affecting other data characteristics. This work is an attempt to provide a review of recent digital image watermarking approaches for forensic applications.

Nowadays, people are shifting from web-based applications to android-based applications to connect people socially. It is important to identify ownership of the messages in android-based applications. Generally, the android-based social media apps (e.g., WhatsApp, etc.) provide end to end encryption to provide security. Therefore, one can never know about the origin of the forwarded information effortlessly. To know the source user, the tracking of each user who forwarded the data is required by the authorized people. But in the presented work, it is possible to know the source/origin of the information without keeping the track record of the people forwarded that data (i.e., image). It saves time and memory. In this proposed work, an image watermarking approach is implemented on the android platform to identify users and apps. The user is identified using the mobile number which is authenticated using OTP verification by Google firebase. The android-based applications (e.g., WhatsApp, etc.) is identified by using orthogonal Walsh codes as application-specific watermarks. Two different approaches are used for watermarking for performance evaluation and experimental results prove the significance of the proposed work.

The use of social media is common in day to day life. People transfer digital data in the form of images, texts or videos to convey the news/information to others. Sometimes, people unintentionally forward the wrong/sensitive information over social media, which can cause negative effects in society. Most of the times this information message is only forwarded, which is being shared by others but the source of that information is the specific user. It is important to find such users, who start spreading this type of digital information over the internet. Such users should be identified and stopped them from doing such activities. A comparative analysis is proposed using two different transforms namely Slantlet transform (SLT) and discrete wavelet transform (DWT) in watermark reconstruction for user identification over the internet. In the proposed robust image watermarking, a 10-digit number is considered and each digit is assigned to a 32 bit randomized Walsh code and 8 additional digits are reserved as padding to obtain the binary watermark. The watermark insertion and extraction process are performed using two different transforms to compare their performance in terms of different watermarking parameters. Although both transforms provide significant results against many signal processing attacks, the experimental evaluation proved that the Slantlet transform gives better results as compared to the DWT.

This paper proposed a transform domain watermarking scheme for ownership and source identification for mobile phone-based messenger applications. In this scheme, text message is used as a watermark signal, which contains user’s 10-digit contact number and secret data representing specific messenger application (i.e. Facebook Messenger, WhatsApp, etc.). Walsh orthogonal codes are used to generate the application-specific watermark signal, which is then merged with user’s cell number to generate complete watermark data. Block-wise slantlet transform is performed on each uniform non-overlapping block of the host image for watermark embedding to achieve robustness against signal processing attacks. Three copies of watermark are inserted to make the scheme more reliable. The correct app identification enables us to know, whether a message is forwarded/not. The correct user identification enables us to know the origin of message. Experimental results signified that the method is robust enough against general signal processing attacks and the objective of the method (successful detection of the user and messenger) is achieved.

This paper proposes a blind image watermarking algorithm based on the transform domain approach. The insertion of watermark is done by replacing the coefficients of DWT-DCT (Discrete Wavelet & Discrete Cosine Transform) transformed host image. Secret keys are used in this method to make it secure against any unauthentic access of watermark as the keys must be known to the user before watermark extractions. The method provides good imperceptibility and also helps to attain good results against various attacks. In this study, peak signal to noise ratio (PSNR) and cross correlation coefficient (CC) are measured to analyse the effect of different attacks on watermarked image signal as well as watermarking quality. Transform domain approach helps this scheme to achieve good robustness against attacks and key based insertion makes this approach secured towards unauthentic access. Proposed method can be used for such image watermarking applications, where good robustness, blindness and security are the prime objectives.