Academia.eduAcademia.edu

Image Forgery Detection

description136 papers
group244 followers
lightbulbAbout this topic
Image Forgery Detection is the process of identifying and analyzing alterations made to digital images, utilizing various techniques to discern authentic content from manipulated or fabricated elements. This field encompasses methods from computer vision, machine learning, and digital forensics to ensure the integrity and authenticity of visual information.
lightbulbAbout this topic
Image Forgery Detection is the process of identifying and analyzing alterations made to digital images, utilizing various techniques to discern authentic content from manipulated or fabricated elements. This field encompasses methods from computer vision, machine learning, and digital forensics to ensure the integrity and authenticity of visual information.

Key research themes

1. How can passive, pixel-level and statistical feature analysis methods detect various types of image forgeries without prior embedded information?

Passive image forgery detection focuses on authenticating digital images without requiring pre-embedded information such as watermarks or signatures. These methods exploit the statistical inconsistencies and artifacts introduced by image manipulation processes, including copy-move and splicing forgeries. Methods under this theme leverage pixel-level analysis, noise estimation, feature point matching, and statistical features to detect tampering even when manipulations are visually imperceptible.

by saba mushtaq and 
1 more
Key finding: This survey categorizes passive authentication techniques into forgery-dependent (targeting specific forgery types like copy-move and splicing) and forgery-independent methods (analyzing artifacts from resampling and lighting... Read more
Key finding: Proposes a novel FFT-DRLBP descriptor that estimates noise patterns using Fast Fourier Transform and encodes discrepancies via Discriminative Robust Local Binary Patterns. This approach captures anomalies in noise patterns... Read more
Key finding: Introduces an advanced keypoint-based copy-move forgery detection approach that overcomes limitations in detecting forgeries in smooth or texture-less regions by employing RootSIFT for keypoint extraction and a multi-scale... Read more
Key finding: Emphasizes the role of physics-based and pixel-level passive techniques that analyze inherent image 'fingerprints' such as sensor noise, chromatic aberrations, and compression artifacts as naturally embedded inconsistencies... Read more
Key finding: Presents an adaptive over-segmentation and feature matching approach combining block-based and keypoint-based forgery detection. By segmenting images into irregular blocks and merging similar feature blocks via color... Read more

2. How do deep learning architectures, including convolutional and recurrent neural networks, enhance accuracy and automation in image forgery detection?

Deep learning models have been increasingly exploited for automated detection and localization of image forgery, offering hierarchical feature extraction and learning capacity to capture complex manipulation patterns. This theme covers the design and evaluation of CNN, ConvLSTM, and hybrid architectures capable of learning discriminative representations from large datasets like CASIA, improving detection across multiple forgery types (copy-move, splicing, retouching) without manual feature engineering.

Key finding: Proposes an integrated deep learning framework combining 2D Convolutional LSTM layers with conventional Convolutional layers evaluated on the CASIA v2.0 dataset. The ConvLSTM (2D) + Conv (2D) architecture outperforms ConvLSTM... Read more
Key finding: Develops a convolutional neural network specifically structured with multiple convolutional and fully connected layers, alongside softmax classification, to differentiate between authentic and tampered images. The model... Read more
Key finding: Introduces an enhanced U-Net architecture for precise segmentation of tampered regions in images obtained from social media platforms, combined with grasshopper optimization for hyperparameter tuning. This method effectively... Read more

3. What multimodal and video-specific techniques effectively detect and localize forgery in videos and deepfake media by exploiting audio-visual inconsistencies and temporal artifacts?

Video forgery detection and deepfake media authentication require analyzing both spatial and temporal domains, including inconsistencies across video frames, motion residuals, as well as multimodal mismatches between audio and visual streams. This theme covers novel approaches integrating motion residual analysis, audio-visual synchronization, ensemble learning, and deep learning methods tailored for video forgery and deepfake detection.

Key finding: Proposes a deep learning based method encompassing motion residual layers to highlight tampering traces, convolutional layers for encoding, and parasitic layers trained via extreme learning theory for classification. The... Read more
Key finding: Develops a multimodal ensemble learning framework combining predictions from separate audio, video, and audiovisual networks alongside a voting module, trained and validated on the FakeAVCeleb dataset. This approach achieves... Read more
by Adil Shah and 
1 more
Key finding: Presents a lipreading-based multimodal detection method that analyzes synchronization mismatches between visual lip movements extracted from the video and lip sequences synthesized from audio using a Wav2Lip model. Using a... Read more
Key finding: Surveys ensemble CNN-based and attention-augmented models for face alteration detection in videos, focusing on the use of EfficientNetB4 and other architectures with attention mechanisms to overcome challenges posed by data... Read more
Key finding: Provides a systematic review of passive video tampering detection techniques especially addressing the challenges unique to video such as compression, high computational complexity, and temporal tampering via frame insertion,... Read more

All papers in Image Forgery Detection

The flipped classroom model has gained recognition as a progressive teaching method aimed at enhancing student engagement and academic success. By moving traditional lecture-based content outside of the classroom, students are encouraged... more
Deepfake or fabricated images have become extremely realistic that it is impossible to simply distinguish a real image from a deepfake one with the naked eye. Technology has evolved greatly in this area that allows people to have easy... more
Image forensic is a field in which the authenticity of an image can be verified along with the traces of image tampering can be detected without the knowledge of any pre-embedded or pre-extracted information. Due to the popularity of this... more
Stroke is one of the leading causes of mortality and disability. It can result in complications like facial paralysis which impair communication and emotional expression. Stroke cases are rising in India due to hypertension, diabetes, and... more
In this paper, a detailed evaluation of multi-scale Weber local descriptors (WLD) based image forgery detection method is presented. Multiscale WLD extracts the features from chrominance components of an image, which usually encode the... more
Copy-Move Forgery Detection (CMFD) method is useful for identifying copy and pasted portions in an image. CMFD has demand in forensic investigation, legal evidence and in many other fields. In this paper, the gists of different newly... more
Feeding difficulties are one of the most significant concerns of children with cerebral palsy. Their poor posture, muscle tone, and movement control make them more susceptible to choking and highly reliant on caregivers during feeding... more
License plate recognition (LPR) has always been one of the crucial predicaments faced due to in numerous reasons such as severe lighting conditions, complex background, unpredicted weather conditions, low light and more. This paper is to... more
Numerous aspects of daily life contribute to societal stability, and the security of people's perceptions of the worldonline is one target of various malicious attacks. Professional forgers can now quickly create copy-move, splice,... more
Digital forensics and computer vision must explore image forgery detection and their related technologies. Image fraud detection is expanding as sophisticated image editing software becomes more accessible. This makes changing photos... more
15. NUMBER OF PAGES 5d. PROJECT NUMBER 5e. TASK NUMBER 5f. WORK UNIT NUMBER 5c. PROGRAM ELEMENT NUMBER 5b. GRANT NUMBER 5a. CONTRACT NUMBER Form Approved OMB NO. 0704-0188 3. DATES COVERED (From -To)
Jeanne Baptiste d’Albert nacque a Parigi nel 1670, figlia di Louis Charles «duc de Luynes et de Chevreuse, pair de France, chevalier des ordres du Roi et grand-fauconnier de France » e della sua seconda moglie Anne de Rohan-Guéméné. Fu... more
A number of methods for altering faces in films have been effectively created and made publicly accessible in recent years (e.g., Face Swap, deepfake, etc.). Using these technologies, it is possible to facilitate face video modifications... more
The Mutual Information is an important statistic which quantifies the relation between two probability distributions in terms of the mutual information content across the two (parallel) linguistic corpuses under consideration. This is a... more
Home invasion crimes in South Korea, though relatively rare, still occur, particularly in urban areas like Seoul, where property crimes such as burglary and robbery are more prevalent. Despite low violent crime rates, the rise in property... more
The development of picture editing software over the past several years has led to the establishment of a topic of active research in the field of digital image fraud detection. Passive forgery detection, or Copy Move Forgery Detection... more
One of the most significant methods utilized in the deep learning approach is text recognition. Text recognition is now a very significant activity that is utilized in many applications of current gadgets to recognize images in a detailed... more
In today's world several image manipulation software's are available. Manipulation of digital images has become a serious problem nowadays. There are many areas like medical imaging, digital forensics, journalism, scientific publications,... more
Philippe Schiesser, Le plomb, métal souvent oublié de la numismatique… Mais ni des faussaires, ni de la fabrication monétaire, Money, it’s a hit, Mélanges sonnants et trébuchants offerts à Jean-Marc Doyen, édité par Pierre Cattelain,... more
The development of picture editing software over the past several years has led to the establishment of a topic of active research in the field of digital image fraud detection. Passive forgery detection, or Copy Move Forgery... more
In the digital age, image manipulation is common, often done before publication on social media. However, this can lead to negative impacts, including visual deception. This research aims to detect splicing type image manipulation using... more
German and English Die sogenannte Anubismaske des Roemer- und Pelizaeus-Museums Hildesheim (Inv. PM 1585) wurde 1910 von Wilhelm Pelizaeus im ägyptischen Antikenhandel erworben und 1911 im Rahmen einer Schenkung der Stadt Hildesheim... more
In this paper, a detailed evaluation of multi-scale Weber local descriptors (WLD) based image forgery detection method is presented. Multiscale WLD extracts the features from chrominance components of an image, which usually encode the... more
The increase in the number of vehicles and the alarming rate of theft and defaulters daily prompts the need for sophisticated matching technology to curb car theft, reduce traffic offenders, and any other anomalies/irregularities... more
The technique of video copy-move forgery (CMF) is commonly employed in various industries; digital videography is regularly used as the foundation for vital graphic evidence that may be modified using the aforementioned method. Recently... more
Thisarticlepresentsanalgorithmforreadingbothsingleandmultipledigitalvideoclocksbyusinga context-awarepixelperiodicitymethodandadeeplearningtechnique.Readingdigitalvideoclocks... more
This article presents an algorithm for reading both single and multiple digital video clocks by using a context-aware pixel periodicity method and a deep learning technique. Reading digital video clocks in real time is a very challenging... more
Surveillance videos provide security and increases work efficiency in places of work and homes. as the most acceptable form of evidence, surveillance videos are now tampered to hide actions or convey wrong information. Researchers have... more
Recent advancements have witnessed the emergence of a software tool that utilizes machine learning techniques to generate audio deepfakes with remarkable realism. These audio manipulations involve recreating voices making it increasingly... more
Steganography is a technology in ‘information hiding’ that allows us to hide a message (text, document, image, audio, video, etc.) in an image or video.LSB technique is very simple and popular data hiding method in which cover image holds... more
Using survival function and transmuted formula to produce lifetime models with application on real data set
The increase in the number of vehicles and the alarming rate of theft and defaulters daily prompts the need for sophisticated matching technology to curb car theft, reduce traffic offenders, and any other anomalies/irregularities... more
Copy-move is one of the most common image tampering method. Many schemes have been proposed to detect and locate the forged regions. However, many existing schemes fail when the copied region is rotated or flipped before being pasted. To... more
This paper is concerned with a vital topic in image processing: color image forgery detection. The development of computing capabilities has led to a breakthrough in hacking and forgery attacks on signal, image, and data communicated over... more
With the massive increase of online content, widespread of social media, the popularity of smartphones, and rise of security breaches, image forensics has attracted a lot of attention in the past two decades alongside the advancements in... more
Nowadays photo manipulation made easier to play with the image files even by a layman. The Combining certain elements to create the unique image that can convince even the most experienced set of eyes. Time to time various detection... more
Abstract: Automatic number plate recognition (ANPR) is a picture processing technology which uses a number (license) plate to spot the vehicle. The main objective is to efficiently design an automatic vehicle identification system by... more
Detection of image forgery is always a crucial factor in image forensic and security applications. Usually this detection is possible with the help of local or global features of an image. We can ensure the credibility of an image with a... more
Background: Cancer is a complex disease which can engages the immune system of the patient. In this regard, determination of distinct immunosignatures for various cancers has received increasing interest recently. However, prediction... more
The evolution of modern cameras, mobile phones equipped with sophisticated image editing software has revolutionized digital imaging. In the process of image editing, contrast enhancement is a very common technique to hide visual traces... more
With the advancement of the multimedia technology, the extensive accessibility of image editing applications makes it easier to tamper the contents of digital images. Furthermore, the distribution of digital images over the open channel... more
This project proposes an efficient realization of deep learning, which monitors the conspicuous behaviour of animals in disaster-prone areas and alerts the authorities just in case of any uncertainty. This project entails monitoring... more
In this article, we propose a novel approach for discerning which scanner has been used to scan a particular document. Its originality relates to a signature extracted in the wavelet domain of the digitized documents where the acquisition... more
Deepfake and manipulated digital photos and videos are being increasingly used in a myriad of cybercrimes. Ransomware, the dissemination of fake news, and digital kidnapping-related crimes are the most recurrent, in which tampered... more
Video-based fire detection (VFD) technologies have received significant attention from both academic and industrial communities recently. However, existing VFD approaches are still susceptible to false alarms due to changes in... more
In today’s digital world because of the widespread availability of software tools and advancement in technology created a way for editing or modifying digital data. Availability of low cost software tools such as Adobe Premiere, Magix... more
This paper is concerned with a vital topic in image processing: color image forgery detection. The development of computing capabilities has led to a breakthrough in hacking and forgery attacks on signal, image, and data communicated over... more
Now a day the automation of technology Video processing tools and techniques are available for altering the Videos for forgery. The modification or changes in current Video is vital to detect since this Video can be used in the... more
Download research papers for free!