Image Forensic Method Research Articles

MMHFND: Fusing Modalities for Multimodal Multiclass Hindi Fake News Detection via Contrastive Learning

Multimodal content contains more deception than unimodal information, causing significant social and economic impacts. Current techniques often focus on single modality, neglecting knowledge fusion. While most studies have concentrated on English fake news detection, this study explores multi-modality for low-resource languages like Hindi. This work introduces the MMHFND model, based on M-CLIP, which uses late fusion for coarse (Fake vs Real) and fine-grained (World vs India vs Politics vs News vs Fact-Check) configurations. We extract deep representations from image and text using image transformer ResNet-50, a BERT-based L3cube-HindRoberta text transformer handling headlines, content, OCR text, and image captions, paired M-CLIP transformers, and an ELA (Error Level Analysis) image forensic method incorporating EfficientNet B0 to analyse multimodal news in Hindi language based on Devanagari script. M-CLIP integrates crossmodal similarity mapping of images and texts with retrieved multimodal features. The extracted features undergo redundancy reduction before being channelled into the final classifier. The MAM (Modality Attention Mechanism) is introduced, which generates weights for each modality individually. The MMHFND model uses a computed modality divergence score to identify dissonance between modalities and a modified contrastive loss on the score. We thoroughly analyse HinFakeNews dataset in a multimodal context, achieving accuracy in coarse and fine-grained configurations. We also undertake an ablation study to evaluate outcomes and explore alternative fusion processes on three different setups. The results show that the MMHFND model effectively detects fake news in Hindi with an accuracy of 0.986, outperforming other existing multimodal approaches.

A survey on digital image forensic methods based on blind forgery detection

A survey on digital image forensic methods based on blind forgery detection

Effective image tampering localization with multi-scale ConvNeXt feature fusion

With the widespread use of powerful image editing tools, image tampering becomes easy and realistic. Existing image forensic methods still face challenges of low generalization performance and robustness. In this letter, we propose an effective image tampering localization scheme based on ConvNeXt encoder and multi-scale Feature Fusion (ConvNeXtFF). Stacked ConvNeXt blocks are utilized as an encoder to capture hierarchical multi-scale features, which are then fused in decoder for locating tampered pixels accurately. Combined loss function and effective data augmentation strategies are adopted to further improve the model performance. Extensive experimental results show that both localization accuracy and robustness of the ConvNeXtFF scheme outperform other state-of-the-art ones. The source code is available at https://github.com/multimediaFor/ConvNeXtFF.

Recaptured image forensics based on local ternary count of high order prediction error

With the advancement in image display technology and the widespread use of various image acquisition devices, recapturing high-quality images from high-fidelity LCD (liquid crystal display) screens becomes very convenient and easy. These recaptured images pose serious threats on current image forensic technology and intelligent recognition systems. In order to solve these security problems, we propose a recaptured image forensics method based on local ternary count (LTC) of high order prediction error. Specifically, given an image, the second-order difference maps on different directions are calculated in HSV color space for the original image and its average pooled version. Then, we calculate the prediction error maps from these second-order difference maps through a linear predictor. Finally, the normalized LTC histograms of all the downsampled versions of the obtained prediction error maps are calculated as features for training and testing. Single database experiments demonstrate that our proposed method performs very similar with the state-of-the-art traditional method on the most difficult-to-detect ICL-COMMSP database with obvious less execution-time and lower feature dimension. Meanwhile, it performs very close to the existing well-performed deep learning based method on all the databases, which verifies the effectiveness of our proposed features. Besides, mixed database experiments verify the superiority on generalization ability of our proposed method. Moreover, it is robust to JPEG compression.

MSRD-CNN: Multi-Scale Residual Deep CNN for General-Purpose Image Manipulation Detection

The authenticity of digital images is a major concern in multimedia forensics due to the availability of advanced photo editing tools/devices. In the literature, several image forensic methods are available to detect specific image processing or editing operations. However, it remains a challenging task to design a universal forensic method that can detect multiple image editing operations. In this paper, a novel Multi-Scale Residual Deep CNN (MSRD-CNN) is designed to learn the image manipulation features adaptively for multiple image manipulation detection. Our network comprises of three stages: pre-processing, hierarchical high-level feature extraction, and classification. Firstly, a multi-scale residual module is employed in pre-processing stage to extract the prediction error or noise features adaptively. Afterwards, the obtained noise features are processed by feature extraction network having multiple Feature Extraction Blocks (FEBs) for the extraction of high-level image tampering features. Lastly, the resultant feature map is provided to the fully-connected dense layer for classification. The experiment results show that our model surpasses the existing schemes even under anti-forensic attacks, when evaluated on large-scale datasets by considering multiple image processing operations. The proposed network provides overall classification accuracies of 97.07% and 97.48% for BOSSBase and Dresden datasets, respectively.

A Document Image Generation Scheme Based on Face Swapping and Distortion Generation

Nowadays, digital images are widely used in various services. The emergence of more and more image editing algorithms has made image forensic approaches to be severely challenged. Driven by the emergence of forged images, more and more image forensic methods are proposed to evaluate the authenticity of digital images. However, in some privacy-related image forensics areas, the scarcity of data affects their development. In this paper, we investigate a document image generation scheme based on face swapping and distortion generation to generate document image databases at a low cost. First, we propose an IDNet for editing face content and text content in digital document images. Second, we propose a Distortion Simulation Network (DSNet) for simulating print-and-scan distortion, and the generated data can be used to study a novel document attack type called recapture attack. Third, we use the database generated by our proposed method to assist in training document image recapture forensic networks. A qualitative comparison with existing methods illustrates that the image content generated by our proposed method maintains more complete semantic information and higher image quality. Quantitative results have confirmed that with the addition of the generated database, the performance of the model improves by about 8% measured in AUC.

Forged Copy-Move Recognition Using Convolutional Neural Network

Due to the extreme robust image editing techniques, digital images are subject to multiple manipulations and decreased costs for digital camera and smart phones. Therefore, image credibility is becoming questionable, specifically when images have strong value, such as news report and insurance claims in a crime court. Therefore, image forensic methods test the integrity of the images by applying various highly technical methods set out in the literature. The present work deals with one important research module is the recognition of forged part that applied on copy move forgery images. Two datasets MICC-F2000 and CoMoFoD are used, these datasets are usually adopted in the field of interest. The module concerned with recognizing which is the source image portion and which is the target one of that already detected. Thus, the two detected tampered parts of the image are recognized the original one from them, the other is then referred as forged or tampered part. The proposed module used the buster net of three neural networks that basically adopted the principle of training by using Convolution Neural Network (CNN) to extract the most important features in the images. The first and second networks are parallel working to detect and identify areas that have been tampered with, and then display them through two masks. While the last network classifier takes a copy of these two catchers to decide which is the source image portion from the two detected ones. The achieved recognition results were about F-score 98.98% even if the forged area is rotated or scaled or both of them. Also, the recognition results of the forged image part was 98% when using images do not contributed in the training phase, which refers to that the proposed module is more confident and reliable.

Diffusion-based image inpainting forensics via weighted least squares filtering enhancement

The rapid development of blockchain technology has greatly changed people’s daily lives from financial market to healthcare field. Today, adulterated images appear in large numbers, causing a serious threat to personal privacy and social stability. At present, the research on blockchain is mainly focused on chain data transmission. Though the blockchain can ensure that the chain data transmission is not tampered, it is difficult to ensure that the data is real when placed on the system initially. In addition, the immutability will be destroyed when 51% attacks occur. Taking this point into consideration, it is necessary to identify the image authenticity on the blockchain. Diffusion-based inpainting is a common method of image tampering. Considering the blurring effect introduced by diffusion-based image inpainting, this paper proposes an image forensics method of diffusion-based image inpainting via weighted least squares filtering enhancement. The texture of the forged image is clear in the untouched regions, and the blurring effect leads to some texture changes in the inpainted regions. Weighted least squares filtering can preserve the texture structure of the untouched regions better and highlight the blurring effect of the inpainted regions. In view of the different reflects of the tampered information in different color channels, weighted least squares filtering is applied to enhance each color channel of the input image, which can capture the impact of image inpainting from multiple perspectives. The experimental results show that the proposed method not only makes up for the deficiency of previous blockchain forensics effectively, but also has better detection performance than the existing work.

Generative Adversarial Ensemble Learning for Face Forensics

The recent advance of synthetic image generation and manipulation methods allows us to generate synthetic face images close to real images. On the other hand, the importance of identifying the synthetic face images increases more and more to protect personal privacy from those. Although some deep learning-based image forensic methods have been developed recently, it is still challenging to distinguish synthetic images generated by recent image generation and manipulation methods such as the deep fake, face2face, and face swap. To resolve this challenge, we propose a novel generative adversarial ensemble learning method. We train multiple discriminative and generative networks based on the adversarial learning. Compared to the conventional adversarial learning, our method is however more focused on improving the discrimination ability rather than image generation one. To this end, we improve the discriminabilty by ensembling outputs from different two discriminators. In addition, we train two generators in order to generate general and hard synthetic images. By ensemble learning of all the generators and discriminators, we improve the discriminators by using the generated synthetic face images, and improve the generators by passing the combined feedback of the discriminators. On the FaceForensics benchmark challenge, we thoroughly evaluate our methods by comparing the recent methods. We also provide the ablation study to prove the effectiveness and usefulness of our method.

A Robust Blind Image Forensic-Detection Method Based On Fuzzy Technique

Blind forensic-investigation in a digital image is a new research direction in image security. It aims to discover the altered image content without any embedded security scheme. Block and key point based methods are the two dispensation options in blind image forensic investigation. Both the techniques exhibit the best performance to reveal the tampered image. The success of these methods is limited due to computational complexity and detection accuracy against various image distortions and geometric transformation operations. This article introduces different blind image tampering methods and introduces a robust image forensic investigation method to determine the copy-move tampered image by means of fuzzy logic approach. Empirical outcomes facilitate that the projected scheme effectively classifies copy-move type of forensic images as well as blurred tampered image. Overall detection accuracy of this method is high over the existing methods.

A multi-purpose image forensic method using densely connected convolutional neural networks

Multi-purpose forensics is attracting increasing attention worldwide. In this paper, we propose a multi-purpose method based on densely connected convolutional neural networks (CNNs) for simultaneous detection of 11 different types of image manipulations. An efficient CNN structure has been specifically designed for forensics by considering vital architecture components, including the number of convolutional layers, the size of convolutional kernels, the nonlinear activations, and the type of pooling layer. The dense connectivity pattern, which has better parameter efficiency than the traditional pattern, is explored to strengthen the propagation of features related to image manipulation detection. When compared with four state-of-the-art methods, our experiments demonstrate that the proposed CNN architecture can achieve better performance in multiple operation detections for different image sizes, especially on small image patches. Consequently, the proposed method can accurately detect local image manipulations. The proposed method can achieve better overall performance when tested on different databases as well as better robustness against JPEG compression even under low-quality JPEG compression.

Recent Advances in Passive Digital Image Security Forensics: A Brief Review

Abstract With the development of sophisticated image editing and manipulation tools, the originality and authenticity of a digital image is usually hard to determine visually. In order to detect digital image forgeries, various kinds of digital image forensics techniques have been proposed in the last decade. Compared with active forensics approaches that require embedding additional information, passive forensics approaches are more popular due to their wider application scenario, and have attracted increasing academic and industrial research interests. Generally speaking, passive digital image forensics detects image forgeries based on the fact that there are certain intrinsic patterns in the original image left during image acquisition or storage, or specific patterns in image forgeries left during the image storage or editing. By analyzing the above patterns, the originality of an image can be authenticated. In this paper, a brief review on passive digital image forensic methods is presented in order to provide a comprehensive introduction on recent advances in this rapidly developing research area. These forensics approaches are divided into three categories based on the various kinds of traces they can be used to track—that is, traces left in image acquisition, traces left in image storage, and traces left in image editing. For each category, the forensics scenario, the underlying rationale, and state-of-the-art methodologies are elaborated. Moreover, the major limitations of the current image forensics approaches are discussed in order to point out some possible research directions or focuses in these areas.

Steganography forensics method for detecting least significant bit replacement attack

Abstract. We present an image forensics method to detect least significant bit replacement steganographyattack. The proposed method provides fine-grained forensics features by using the hierarchical structurethat combines pixels correlation and bit-planes correlation. This is achieved via bit-plane decomposition anddifference matrices between the least significant bit-plane and each one of the others. Generated forensics fea-tures provide the susceptibility (changeability) that will be drastically altered when the cover image is embeddedwithdatatoformastegoimage.Wedevelopedastatisticalmodelbasedontheforensicsfeaturesandusedleastsquare support vector machine as a classifier to distinguish stego images from cover images. Experimentalresults show that the proposed method provides the following advantages. (1) The detection rate is noticeablyhigher than that of some existing methods. (2) It has the expected stability. (3) It is robust for content-preservingmanipulations, such as JPEG compression, adding noise, filtering, etc. (4) The proposed method providessatisfactory generalization capability.

Passive Image Forensic Method to detect Copy Move Forgery in Digital Images

Tampering in digital images has become very easy due to the availability of advanced image editing softwares to the users. Images are being tampered in a very efficient manner without leaving any visual clue. As a consequence, the content of digital images cannot be taken as for granted. There are various types of image tampering techniques. One of the most common tampering techniques is copy-move forgery. In copy-move forgery one part of an image is copied and pasted in another part of the same image. In this paper, the passive image forensic method is presented to detect copy move forgery in digital images. The proposed method is based on SURF (Speed Up Robust Features) algorithm. In this method the features are extracted and their descriptors are obtained by SURF algorithm and the Nearest Neighbor approach is used for feature matching to identify the copy move forgery in digital images. This detection method is found to be rotation and scale invariant and is robust enough to noise, jpeg compression and blurring. Multiple copy move forgery is also detected by this method.

Passive detection of copy-paste forgery between JPEG images

A blind digital image forensic method for detecting copy-paste forgery between JPEG images was proposed. Two copy-paste tampering scenarios were introduced at first: the tampered image was saved in an uncompressed format or in a JPEG compressed format. Then the proposed detection method was analyzed and simulated for all the cases of the two tampering scenarios. The tampered region is detected by computing the averaged sum of absolute difference (ASAD) images between the examined image and a resaved JPEG compressed image at different quality factors. The experimental results show the advantages of the proposed method: capability of detecting small and/or multiple tampered regions, simple computation, and hence fast speed in processing.

Image Forensics Using Generalised Benford's Law for Improving Image Authentication Detection Rates in Semi-Fragile Watermarking

In the past few years, semi-fragile watermarking has become increasingly important to verify the content of images and localise the tampered areas, while tolerating some non-malicious manipulations. In the literature, the majority of semi-fragile algorithms have applied a predetermined threshold to tolerate errors caused by JPEG compression. However, this predetermined threshold is typically fixed and cannot be easily adapted to different amounts of errors caused by unknown JPEG compression at different quality factors (QFs). In this paper, the authors analyse the relationship between QF and threshold, and propose the use of generalised Benford’s Law as an image forensics technique for semi-fragile watermarking. The results show an overall average QF correct detection rate of approximately 99%, when 5%, 20% and 30% of the pixels are subjected to image content tampering and compression using different QFs (ranging from 95 to 65). In addition, the authors applied different image enhancement techniques to these test images. The proposed image forensics method can adaptively adjust the threshold for images based on the estimated QF, improving accuracy rates in authenticating and localising the tampered regions for semi-fragile watermarking.