Search

Rationale and Objectives Appropriate image manipulation of angiographic image display systems during interventional radiology is performed by radiological technologists and/or nurses given instructions from radiologists. However, appropriate images might not be displayed because of communication errors. Therefore, we developed a manipulation system that uses an eye tracker. The study aimed to determine if an angiographic image display system can be manipulated as well by using an eye tracker as by using a mouse. Materials and Methods An angiographic image display system using an eye tracker to calculate the gaze position on the screen and state of fixation was developed. Fourteen radiological technologists participated in an observer study by manipulating 10 images for each of 5 typical cases frequently performed in angiography, such as renal tumor, cerebral aneurysm, liver tumor, uterine bleeding, and hypersplenism. We measured the time from the start to the end of manipulating a series of images required when using the eye tracker and the conventional mouse. In this study, the statistical processing was done using Excel and R and R studio. Results The average time required for all observers for completing all cases was significantly shorter when using the eye tracker than when using the mouse (10.4 ± 2.1 s and 16.9 ± 2.6 s, respectively; p Conclusion Radiologists were able to manipulate an angiographic image display system directly by using the newly developed eye tracker system without touching contact devices, such as a mouse or angiography console. Therefore, communication error could be avoided.

Purpose A procedure is needed for bodies with disfiguring injuries to the face and the use of their portrait for visual identification. Method We present the application of a simple image processing procedure, otherwise known as ”bubbling,” which is based on the concept of ”perceptual filling-in,” to images for visual identification in the forensic context. The method is straight forward and can be performed using readily available software and hardware.. Results The method is demonstrated and examples are shown. The visual recognition of known persons using “bubbled” images was successfully tested. Conclusion The “bubbling” procedure for visual identification enhancement is quick and straightforward and may be attempted before escalating to more involved identification methods and procedures.

A new style transfer-based image manipulation framework combining generative networks and style transfer networks is presented in this paper. Unlike conventional style transfer tasks, we tackle a new task, text-guided image manipulation. We realize style transfer-based image manipulation that does not require any reference style images and generate a style image from the user’s input sentence. In our method, since an initial reference input sentence for a content image can automatically be given by an image-to-text model, the user only needs to update the reference sentence. This scheme can help users when they do not have any images representing the desired style. Although this text-guided image manipulation is a new challenging task, quantitative and qualitative comparisons showed the superiority of our method.

BACKGROUND The aim of this study was to determine if image enhancement improves a clinician's ability to identify the presence of calculus on digital radiographs. METHODS Seventy-one hopeless teeth were collected from 34 patients. Teeth were stained with 1% methylene blue, the largest interproximal calculus deposit was scored, and photographs of each interproximal root surface were taken. The surface area of calculus deposit was determined as a percentage of the total interproximal root surface area. Digital radiographs of teeth taken before extraction were modified using the following enhancements: auto-contrast, emboss, invert, and sharpen. Radiographic presence of calculus was determined by two examiners. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated for each examiner and enhancement. A receiver operating characteristic curve was used to compare differences between the image enhancements in the detection of dental calculus. The kappa statistic was used to compare ratings between examiners. RESULTS None of the enhanced images were statistically superior to original images in identifying radiographic calculus (P > 0.05). The average sensitivity of digital radiography was 50%, average specificity was 82.2%, PPV was 94%, and NPV 23.2%. A threshold of >30% of interproximal root surface covered with calculus and increasing size of deposits were associated with improved detection (P < 0.05). CONCLUSIONS Digital enhancements do not significantly improve radiographic detection of dental calculus. As area of calculus on the root surface and size of calculus deposits increased, sensitivity of detection also increased.

Given a potentially manipulated probe image, provenance analysis aims to find all images derived from the probe (offsprings) and all images from which the probe is derived (ancestors) in a large dataset (provenance filtering), and reconstruct the manipulation history with the retrieved images (provenance graph building). In this paper, we address two major challenges in provenance analysis, retrieving the source image of the small regions that are spliced into the probe image, and, detecting source images within the search results. For the former challenge, we propose to detect spliced regions by pairwise image comparison and only use local features extracted from the spliced region to perform an additional search. This removes the influence of the background and greatly improves the recall. For the latter, we propose to learn a pairwise ancestor-offspring detector and use it jointly with a holistic image manipulation detector to identify the source image. The proposed provenance analysis system has performed remarkably in evaluations using comprehensive provenance datasets. It's the winning solution for NIST Media Forensics Challenge (MFC) in 2018, 2019 and 2020. In MFC 2019, our provenance results achieved a 12% improvement in filtering and a 20% gain in oracle provenance graphs building over the alternative methods. In the real-world Reddit dataset, the edge overlap between our reconstructed provenance graphs and the ground-truth graphs is 5 times better than the state-of-the-art system.

Deepfakes are a form of human image synthesis where an existing picture or image is superimposed into a video to change the identity of those depicted in the video. The technology relies on machine...

Image forgery detection using traditional algorithms takes much time to find forgeries. The new emerging methods for the detection of image forgery use a deep neural network algorithm. A hybrid deep learning (DL) and machine learning-based approach is used in this study for passive image forgery detection. A DL algorithm classifies images into the forged and not forged categories, whereas colour illumination localises forgery. The simulated results are compared to other algorithms on public datasets. The simulated results achieved 99% accuracy for CASIA1.0, 98% accuracy for CASIA2.0, 98% accuracy for BSDS300, 97% accuracy for DVMM, and 99% accuracy for CMFD image manipulation dataset.

Misconduct in terms of manipulation of images has become an increasingly serious issue for the scientific community, especially in biomedicine. Such misconduct takes many forms, the major categories being falsified or fabricated images, manipulated images, and plagiarized images. Different tools and techniques are briefly described to help authors and editors in detecting such misconduct, and guidance is offered on appropriate use of images under different situations. More specifically, CrossRef Similarity Check, Motuin, and Droplets are proposed as the tools of choice for detecting similarity between images and their possible manipulation.

The monitoring of the intake os radionuclides by workers and public individuals, using in vivo measurements, requires the application of calibration factors obtained with physical-anthropomorphic phantoms. In the case of 131I, the lack of anatomical realism of the phantom might impair the quality and reliability of the monitoring result. Thus, in order to develop a thyroid phantom by means of 3D printing, images of the voxelized models provided in ICRP 110 were used. The superposition of the images of the original model and the treated model demonstrate the maintenance of morphological characteristics. Therefore, image manipulation techniques applied in this work aimed to smooth the sharp angles of the original image and the prototype printing were considered effective.

State-of-the-art multi-stream methods for image manipulation detection suffer from gaps between features. Moreover, the rapid development of GANs makes it an emerging method of image tampering. However, existing natural scene image tampered datasets are limited to manual tampering. In this paper, we address these two issues. Firstly, we propose a novel two-stream multi-modality image manipulation detection model (MM-net) that abandons the way of fusion. The main idea of the proposed model is to exploit one stream to guide the learning of the other stream through attention mechanism. Therefore, our approach enjoys the benefits of the multi-stream methods while avoiding the semantic gaps caused by bridging gaps between different streams. Secondly, we build the first tampered dataset of natural images based on GANs, pushing manipulation detection toward more realistic and challenging scenarios. Extensive experimental results demonstrate that our model outperforms state-of-the-art approaches on both manually and GANs tampered images.

We review the phenomenon of deepfakes, a novel technology enabling inexpensive manipulation of video material through the use of artificial intelligence, in the context of today’s wider discussion on fake news. We discuss the foundation as well as recent developments of the technology, as well as the differences from earlier manipulation techniques and investigate technical countermeasures. While the threat of deepfake videos with substantial political impact has been widely discussed in recent years, so far, the political impact of the technology has been limited. We investigate reasons for this and extrapolate the types of deepfake videos we are likely to see in the future.

International audience; A novel two-stage approach is proposed for image manipulation and generation. User-interactive image deformation is performed through editing of contours. This is performed in the latent edge space with both color and gradient information. The output of editing is then fed into a multi-scale representation of the image to recover quality output. The model is flexible in terms of transferability and training efficiency.

A growing number of commercial satellite companies provide easily accessible satellite imagery. Overhead imagery is used by numerous industries including agriculture, forestry, natural disaster analysis, and meteorology. Satellite images, just as any other images, can be tampered with image manipulation tools. Manipulation detection methods created for images captured by "consumer cameras" tend to fail when used on satellite images due to the differences in image sensors, image acquisition, and processing. In this paper we propose an unsupervised technique that uses a Vision Transformer to detect spliced areas within satellite images. We introduce a new dataset which includes manipulated satellite images that contain spliced objects. We show that our proposed approach performs better than existing unsupervised splicing detection techniques.

Introduction Significant advances in three-dimensional (3D) imaging technology have allowed for the incorporation of 3D digital displays into medical and surgical devices. Despite initial adoption of the NGENUITY® 3D Visualization System in vitreoretinal surgery, there are limited publications regarding its use. The generally accepted main benefits include improved ergonomics, enhanced surgical team communication and education, reduced retinal phototoxicity, increased depth of field, and display image manipulation. Despite these potential benefits, many retina specialists have questioned its universal applicability to a wide variety of vitreoretinal surgeries. Objective To report on the variety of indications and surgical efficacy of the NGENUITY® 3D Visualization System in vitreoretinal surgery via a review of surgical experience at two vitreoretinal practices in both the academic and community settings. Methods A retrospective review was conducted of consecutive surgical cases performed on the NGENUITY® 3D Visualization System at Massachusetts Eye and Ear Infirmary and Florida Retina Institute from June 1st, 2017 to November 1st, 2018. Age, presenting diagnosis, surgical procedure, and intraoperative details were recorded. Results 272 vitreoretinal surgeries on the Alcon NGENUITY® 3D Visualization System were identified between June 1st, 2017 and November 1st, 2018 at the participating institutions. A detailed breakdown of the indications for surgery and related procedures is reported. During all 272 cases on the 3D digital system, there were no complications attributed to the visualization system. Conclusion This series illustrates the diversity of vitreoretinal surgeries that can be performed on this system without compromising surgical viewing or increasing surgical complications. The Alcon NGENUITY® 3D Visualization System possesses favorable ergonomics, illumination levels, depth of field, display filters, and trainee experience.

This work discusses the development of an augmented reality (AR) interface that allows users to manipulate 3D anatomical structures (segmented and reconstructed from real medical studies) through haptic interactions by means of a five-degrees-of-freedom hand-exoskeleton. The exoskeleton allows the hand to move throughout its nominal range of motion. Our interface allows users to manipulate virtual objects more naturally thanks to its three points of contact, unlike most commercial haptic systems, which only have one point of contact. The preliminary results indicated a proper 3D segmentation and reconstruction of the anatomical structures of interest, as well as an adequate interaction between the haptic exoskeleton and the AR interface.

Facial image manipulation is a particular instance of digital image tampering, which is done by compositing a region from one facial image into another facial image. Fake images generated by facial image manipulation now spread like wildfire on news websites and social networks, and are considered the greatest threat to press freedom. Previous research relied heavily on handcrafted features to analyze tampered regions which were inefficient and time-consuming. This paper introduces a framework that accurately detects manipulated face image using deep learning approach. The original contributions of this paper include (1) a customized convolutional neural network model for Manipulated Face (MANFA) identification; it contains several convolutional layers that effectively extract features of multi-levels of abstraction from a tampered region. (2) A hybrid framework (HF-MANFA) that uses Adaptive Boosting (AdaBoost) and eXtreme Gradient Boosting (XGBoost) to deal with the imbalanced dataset challenge. (3) A large manipulated face dataset that is manually collected and validated. The results from various experiments proved that proposed models outperformed existing expert and intelligent systems which were usually used for the manipulated face image detection task in terms of area under the curve (AUC), computational complexity, and robustness against imbalanced datasets. As a result, the presented framework will motivate the finding of a more powerful altered face images detection method and encourages the integration of the proposed model in applications that have to deal with manipulated images regularly.

In forensic intelligence-gathering, footprints have been shown to be valued evidence found at crime scenes. Forensic podiatrists and footprint examiners use a variety of techniques for measuring footprints for comparison of the crime scene evidence with the exemplar footprints. This study examines three different techniques of obtaining two-dimensional linear measurement data of dynamic bare footprints. Dynamic bare footprints were gathered from 50 students from a podiatric medical school using the Identicator® Inkless Shoe Print Model LE 25P system. After obtaining 100 bilateral footprints from the participants, the quantitative measurement data were collected by using three different measurement techniques: (i) a manual technique using a ruler (direct technique); (ii) an Adobe® Photoshop® technique; and (iii) a GIMP (GNU Image Manipulation Program) technique. The seven Reel linear measurement methodology was used for producing measurements using these three techniques. This study showed that all the mean bare footprint measurements on the right and left feet obtained using the direct technique were larger than those obtained using GIMP and Adobe® Photoshop® images. Differences were also observed in measurements produced using GIMP software and Photoshop images. However, the differences observed in the three techniques used for bare footprint measurements were not found to be statistically significant. The study concludes that there are no significant differences between the three measurement techniques when applied to two-dimensional bare footprints using the Reel method. It further concluded that any of these measurement techniques can be used when employing the Reel methodology for footprint analysis without significant difference.

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.

This position paper addresses three areas in which photographic images are used and abused: Journalism, Advertising, and Public Relations. These areas do not use images for aesthetic purposes only; for these businesses, the images are more than mental impressions. In these three industries, the messages they deliver through the images have little to do with the objective reality. All three use images to deliver an opinion in the name of journalism, distorted needs, benefits, and solutions in advertising, and put spins on facts to present favorable images of their clients in Public Relations. Historically, all three have engaged in the unethical use of the concept “image making” and of manipulation of actual images.

Encryption-then-Compression (EtC) systems have been proposed to securely transmit images through an untrusted channel provider. In this study, EtC systems were applied to social media like Twitter that carry out image manipulations. The block scrambling-based encryption schemes used in EtC systems were evaluated in terms of their robustness against image manipulation on social media. The aim was to investigate how five social networking service (SNS) providers, Facebook, Twitter, Google+, Tumblr and Flickr, manipulate images and to determine whether the encrypted images uploaded to SNS providers can avoid being distorted by such manipulations. In an experiment, encrypted and non-encrypted JPEG images were uploaded to various SNS providers. The results show that EtC systems are applicable to the five SNS providers., Comment: To be appeared in IEICE Transactions on Information and Systems, January 2019

Aim/Purpose: In this paper, we address the theoretical challenges today’s scientific community faces to precisely draw lines between true and false pictures. In particular, we focus on problems related to the hidden wonders of science and the shiny images produced for scientific papers or to appeal to wider audiences. Background: As rumors (hoaxes) and false news (fake news) explode across society and the current network, several initiatives using current technology have been launched to study this phenomena and limit the social impact. Over the last two decades, inappropriate scientific behavior has raised more questions about whether some scientific images are valid. Methodology: This work is not about analyzing whether today’s images are objective. Instead, we advocate for a general approach that makes it easier to truly believe in all kinds of knowledge, scientific or otherwise (Goldman, 1967; Goldman, & Olson, 2009). This need to believe is closely related to social order (Shapin, 1994). Contribution: We conclude that we must ultimately move away from older ideas about truth and objectivity in research to broadly approach how science and knowledge are represented and move forward with this theoretical approach when communicating science to the public. Findings: Contemporary visual culture suggests that our world is expressed through images, which are all around us. Therefore, we need to promote the reliability of scientific pictures, which visually represent knowledge, to add meaning in a world of complex high-tech science (Allamel-Raffin, 2011; Greenberg, 2004; Rosenberger, 2009). Since the time of Galileo, and today more than ever, scientific activity should be understood as knowledge produced to reveal, and therefore inform us of, (Wise, 2006) all that remains unexplained in our world, as well as everything beyond our senses. Recommendation for Researchers: In journalism, published scientific images must be properly explained. Journalists should tell people the truth, not fake objectivity. Today we must understand that scientific knowledge is mapped, simulated, and accessed through interfaces, and is uncertain. The scientific community needs to approach and explain how knowledge is represented, while paying attention to detail. Future Research: In today’s expanding world, scientific research takes a more visual approach. It is important for both the scientific community and the public to understand how the technologies used to visually represent knowledge can account for why, for example, we know more about electrons than we did a century ago (Arabatzis, 1996), or why we are beginning to carefully understand the complexities and ethical problems related to images used to promote knowledge through the media (see, i.e., López-Cantos, 2017).

Deepfake image manipulation has become a serious security threat to the social network. Currently, there are limited studies on protective methods that are against Deepfake image manipulation. To tackle this problem, we here propose an adversary attack based smart watermark model, which adds unperceptive watermarks to images so that the images become adversary examples to Deepfake models. When the Deepfake manipulates these watermarked images, the manipulated images become blur. The manipulation thus can be easily recognized by human and machines. Our experiments have shown that our model outperforms the SOTA and can be used to effectively prevent Deepfake manipulation.

In this paper, we propose a novel CycleGAN without checkerboard artifacts for counter-forensics of fake-image detection. Recent rapid advances in image manipulation tools and deep image synthesis techniques, such as Generative Adversarial Networks (GANs) have easily generated fake images, so detecting manipulated images has become an urgent issue. Most state-of-the-art forgery detection methods assume that images include checkerboard artifacts which are generated by using DNNs. Accordingly, we propose a novel CycleGAN without any checkerboard artifacts for counter-forensics of fake-mage detection methods for the first time, as an example of GANs without checkerboard artifacts.

In the past years, numerous methods have been introduced to reliably detect digital face image manipulations. Lately, the generalizability of these schemes has been questioned in particular with respect to image post-processing. Image compression represents a post-processing which is frequently applied in diverse biometric application scenarios. Severe compression might erase digital traces of face image manipulation and hence hamper a reliable detection thereof. In this work, the effects of image compression on face image manipulation detection are analyzed. In particular, a case study on facial retouching detection under the influence of image compression is presented. To this end, ICAO-compliant subsets of two public face databases are used to automatically create a database containing more than 9,000 retouched reference images together with unconstrained probe images. Subsequently, reference images are compressed applying JPEG and JPEG 2000 at compression levels recommended for face image storage in electronic travel documents. Novel detection algorithms utilizing texture descriptors and deep face representations are proposed and evaluated in a single image and differential scenario. Results obtained from challenging cross-database experiments in which the analyzed retouching technique is unknown during training yield interesting findings: (1) most competitive detection performance is achieved for differential scenarios employing deep face representations; (2) image compression severely impacts the performance of face image manipulation detection schemes based on texture descriptors while methods utilizing deep face representations are found to be highly robust; (3) in some cases, the application of image compression might as well improve detection performance., to appear in IET Biometrics

It has been suggested that luminance edges in retinal images are potential cues for glossiness perception, particularly when the perception relies on low-luminance specular regions. However, a previous study has shown only statistical correlations between luminance edges and perceived glossiness, not their causal relations. Additionally, although specular components should be embedded at various spatial frequencies depending on the micro-roughness on the object surface, it is not well understood what spatial frequencies are essential for glossiness perception on objects with different micro-roughness. To address these issues, we examined the impact of a sub-band contrast enhancement on the perceived glossiness in the two conditions of stimuli: the Full condition where the stimulus had natural specular components and the Dark condition where it had specular components only in dark regions. Object images with various degrees of surface roughness were generated as stimuli, and their contrast was increased in various spatial-frequency sub-bands. The results indicate that the enhancement of the sub-band contrast can significantly increase perceived glossiness as expected. Furthermore, the effectiveness of each spatial frequency band depends on the surface roughness in the Full condition. However, effective spatial frequencies are constant at a middle spatial frequency regardless of the stimulus surface roughness in the Dark condition. These results suggest that, for glossiness perception, our visual system depends on specular-related information embedded in high spatial frequency components but may change the dependency on spatial frequency based on the surface luminance to be judged.

The recent and rapid growth in image manipulation techniques raises grave concerns with respect to document forgery. It also diminishes trust in the data available publicly and can cause harm by spreading false information. In this paper, the problem of facial forgery by training a neural network to distinguish between real and fake faces has been addressed. The convolutional neural network has been trained with the help of transfer learning, which allows us to utilize computer vision models like ResNet and AlexNet, which were pre-trained on massive datasets. The network was evaluated on a dataset of 2041 images provided by the computational Intelligence and Photography Lab at the Department of Computer Science, Yonsei University. Results show that ResNet-152 is able to detect facial forgery with an accuracy of up to 76.79%.

With the development of information security, localization of image manipulations havs become a hot topic. In this paper, a hybrid loss network is proposed for the manipulated image forensics. First, the patch prediction module extracts corresponding features representing the discrepancy of the tampered region boundaries. Then, these features are constrained by the Pixel Normalization, thereby improving the classification performance for the tampered patches. Finally, multi-scale patch prediction masks and semantic information are fused to segment out the tampered regions. The experiments demonstrate the proposed model can achieve high performance.

Satellite imagery is becoming increasingly available due to a large number of commercial satellite companies. Many fields use satellite images, including meteorology, forestry, natural disaster analysis, and agriculture. These images can be changed or tampered with image manipulation tools causing issues in applications using these images. Manipulation detection techniques designed for images captured by “consumer cameras” tend to fail when used on satellite images. In this paper we propose a supervised method, known as Nested Attention U-Net, to detect spliced areas in the satellite images. We introduce three datasets of manipulated satellite images that contain objects generated by a generative adversarial network (GAN). We test our approach and compare it to existing supervised splicing detection and segmentation techniques and show that our proposed approach performs well in detection and localization.

Cross-sectional imaging techniques produce a two-dimensional representation of a section of human anatomy. Further mathematical manipulation of these images can create additional images that may better demonstrate anatomic relationships or make pathology more conspicuous. There is a wide array of techniques that can manipulate medical images in this manner; they are globally referred to as “post-processing.”

SinGAN shows impressive capability in learning internal patch distribution despite its limited effective receptive field. We are interested in knowing how such a translation-invariant convolutional generator could capture the global structure with just a spatially i.i.d. input. In this work, taking SinGAN and StyleGAN2 as examples, we show that such capability, to a large extent, is brought by the implicit positional encoding when using zero padding in the generators. Such positional encoding is indispensable for generating images with high fidelity. The same phenomenon is observed in other generative architectures such as DCGAN and PGGAN. We further show that zero padding leads to an unbalanced spatial bias with a vague relation between locations. To offer a better spatial inductive bias, we investigate alternative positional encodings and analyze their effects. Based on a more flexible positional encoding explicitly, we propose a new multi-scale training strategy and demonstrate its effectiveness in the state-of-the-art unconditional generator StyleGAN2. Besides, the explicit spatial inductive bias substantially improve SinGAN for more versatile image manipulation., paper with appendix, project page: https://nbei.github.io/gan-pos-encoding.html

Scientific fraud by image duplications and manipulations within western blot images is a rising problem. Currently, problematic western blot images are mainly detected by checking repeated bands or through visual observation. However, the completeness of the above methods in detecting problematic images has not been demonstrated. Here we show that Generative Adversarial Nets (GANs) can generate realistic western blot images that indistinguishable from real western blots. The overall accuracy of researchers for identifying synthetic western blot images is 0.52, which almost equal to blind guess (0.5). We found that GANs can generate western blot images with bands of the expected lengths, widths, and angles in desired positions that can fool researchers. For the case study, we find that the accuracy of detecting the synthetic western blot images is related to years of researchers performed studies relevant to western blots, but there was no apparent difference in accuracy among researchers with different academic degrees. Our results demonstrate that GANs can generate fake western blot images to fool existing problematic image detection methods. Therefore, more information is needed to ensure that the western blots appearing in scientific articles are real. We argue to require every western blot image to be uploaded along with a unique identifier generated by the laboratory machine and to peer review these images along with the corresponding submitted articles, which may reduce the incidence of scientific fraud.

A large vascular disorders such as stenosis, aneurysm and malformations, which involve different anatomical locations, are detected and handled through a set of techniques for imaging them. The need to diagnose and manage vascular disorders early has now contributed to the creation of numerous techniques in vascular imaging. Image manipulation plays an significant part of medicine’s study of photographs from different methods for the treatment of eyes. The goal of the novel method for enhancing visualization by using hessian-based filters is to highlight the secret vessels, to improve angiographies as well as the possible pathological locations. The pictures found come from the computed tomography and retinal data. The goal of the novel method for enhancing visualization by using hessian-based filters is to highlight the secret vessels, to improve angiographies as well as the possible pathological locations. The pictures found come from the computed tomography and retinal data. The novel upgrading feature suggested has many applications such as retinal vasculatures, neck, lung and fundus, but only retinal and cerebral vasculatures are taken into account.

Satellite imagery is becoming increasingly accessible due to the growing number of orbiting commercial satellites. Many applications make use of such images: agricultural management, meteorological prediction, damage assessment from natural disasters, or cartography are some of the examples. Unfortunately, these images can be easily tampered and modified with image manipulation tools damaging downstream applications. Because the nature of the manipulation applied to the image is typically unknown, unsupervised methods that don't require prior knowledge of the tampering techniques used are preferred. In this paper, we use ensembles of generative autoregressive models to model the distribution of the pixels of the image in order to detect potential manipulations. We evaluate the performance of the presented approach obtaining accurate localization results compared to previously presented approaches.

Image manipulation detection has evolved as a very challenging problem in the field of multimedia forensics. Object removal from the image is one of the most used manipulation operations in the image. Object removal accomplished by exemplar-based image inpainting makes image visually pleasing and physically plausible without any noticeable trace. So, detection of inpainted region has always been a challenge for researchers. In this paper, we have proposed a semantic segmentation-based approach to detect the inpainted region in an image. This is a deep learning-based approach which uses fully convolutional network. We have achieved a remarkable result in the detection of inpainted region. For training and validation of our results, we have created inpainted image database. We have evaluated our results on different performance metrics.

The aim of the present work is to assess the effect of window width and window level selection on feature size in a 2D displayed multimodality images. DICOM images of CT, MR and PET were retrieved and retrospectively analyzed using a local PACS system running an image manipulation software AGFA NX v2.0 that supports several windows for each image modality. The images analyzed were phantom based using standardized NEMA ICE body phantom that consist of six fillable variable spheres sizes. A NOVA t-test was used for statistical difference studies. Using bone and abdomen/ pelvis CT windows resulted in a sphere underestimation of up to 3 mm. While using T1 and T2 MRI windows yielded an underestimation of up to 3.5 mm especially in the small spheres. Significant spheres size differences were seen with PET windows shown an overestimation of up to 7 mm using CE equal and thermal windows. Improper selection of windowing while viewing 2D displayed images can yield considerable inconsistency leading to unreliable measurement.

Deep image-to-image translation networks have shown satisfying results on domain adaptive image generation tasks. In this paper, we make further progress on a new task of cross-domain image manipulation. We aim at generating a manipulated image that has a specified factor of the source-domain image and other factors of the target-domain image. We present a novel cross-domain disentangle network, which is capable of learning disentangled representation from data across multiple domains. This network extracts the latent representation of images on source and target domains, and disentangles the specified and unspecified factors of latent representation simultaneously on both domains. Image manipulation within and across domain can be realized by manipulating on the latent space. Empirical experiments show that our method is able to generate convincing intra-domain and cross-domain manipulated images.

Protecting image manipulation detectors against perfect knowledge attacks requires the adoption of detector architectures which are intrinsically difficult to attack. In this paper, we do so, by exploiting a recently proposed multiple-classifier architecture combining the improved security of 1-Class (1C) classification and the good performance ensured by conventional 2-Class (2C) classification in the absence of attacks. The architecture, also known as 1.5-Class (1.5C) classifier, consists of one 2C classifier and two 1C classifiers run in parallel followed by a final 1C classifier. In our system, the first three classifiers are implemented by means of Support Vector Machines (SVM) fed with SPAM features. The outputs of such classifiers are then processed by a final 1C SVM in charge of making the final decision. Particular care is taken to design a proper strategy to train the SVMs the 1.5C classifier relies on. This is a crucial task, due to the difficulty of training the two 1C classifiers at the front end of the system. We assessed the performance of the proposed solution with regard to three manipulation detection tasks, namely image resizing, contrast enhancement and median filtering. As a result, the security improvement allowed by the 1.5C architecture with respect to a conventional 2C solution is confirmed, with a performance loss in the absence of attacks that remains at a negligible level., 27 pages, 9 figures, Submitted to the An International Journal Multimedia Tools and Applications-Springer