Block-based Technique Research Articles

A robust detection and localization technique for copy-move forgery in digital images

Image forensic analysis becomes a major role in the field of digital image security due to tampering and forgery. The image forgery violates the authenticity and ownership of digital images. Copy-move forgery considers a significant kind of image forensic analysis algorithm. In this kind, the forger copies a part of an original image and then pastes it into the selected position from the same image. The purpose of forgery is to hide or highlight a specific region of the original image. To detect copy-move forgery, there are two traditional techniques: block-based and keypoint-based. The main drawback of the keypoint-based technique is the insufficient features for the small and flat regions, which causes undetected forgery. In contrast, the block-based technique has intensive processing. Therefore, this paper proposes a robust scheme that overcomes the drawbacks of the above techniques and maintains their advantages. This scheme adopts three connected stages, the first detects the initial duplicated regions using the SURF-HOG detector and descriptor. Subsequently, the second stage localizes the primary matched regions by SLIC segmentation and then selects the suspicious neighbor regions to be combined with primary regions to obtain the active regions. In the third stage, the block-based technique adopts overlapping Zernike moments to extract sufficient key points from the produced active regions. In the final stage, the duplicated regions are classified into authentic or forged regions. The proposed scheme provides not only forgery detection but also localization and recognition for the duplicated regions. The experimental results show that the proposed scheme is fast and has high accuracy for forgery detection and localization, at least 93.75, and 7.25 in terms of True Positive and False Positive Rates. Moreover, the scheme has high robustness under various conditions and attacks such as geometric transformation attacks and compound photometric attacks. The proposed scheme can be used in sensitive applications such as cybercrime detection and adopted as evidence in the courts.

Error concealment Pada Transmisi Video H.264/AVC dengan Frequency Selective Extrapolation

Video transmission is encoded by block-based techniques such as MPEG in sensitive environment. It’s very susceptible to noise which can cause block losses and even missing frames during transmission. A technique to overcome this problem is applied Error concealment (EC) on the decoder. There are two approaches used in this EC, spacial EC which utilizes information around the image and temporal EC by utilizing motion information. The method of EC that used in this research is Frequency Selective Extrapolation (FSE). The video is encoded by the H.264 / AVC standard (MPEG-Part10). The isolated block losses was added to the video as an error simulation, then applied the EC-FSE method on the decoder. There are two method of EC-FSE used, FSE-2D (two dimensions) and FSE-3D (three dimensions). The measurement results were observed by the PSNR and MOS values. The simulation results show that H.264 / AVC video concealed with FSE-3D has a better performance than FSE-2D.

Compression effects and scene details on the source camera identification of digital videos

The continuous growth of technologies like 4G or 5G has led to a massive use of mobile devices such as smartphones and tablets. This phenomenon, combined with the fact that people use mobile phones for a longer period of time, results in mobile phones becoming the main source of creation of visual information. However, its reliability as a true representation of reality cannot be taken for granted due to the constant increase in editing software. This makes it easier to alter original content without leaving a noticeable trace in the modification. Therefore, it is essential to introduce forensic analysis mechanisms to guarantee the authenticity or integrity of a certain digital video, particularly if it may be considered as evidence in legal proceedings. This paper explains the branch of multimedia forensic analysis that allows to determine the identification of the source of acquisition of a certain video by exploiting the unique traces left by the camera sensor of the mobile device in visual content. To do this, a technique that performs the identification of the source of acquisition of digital videos from mobile devices is presented. It involves 3 stages: (1) Extraction of the sensor fingerprint by applying the block-based technique. (2) Filtering the strong component of the PRNU signal to improve the quality of the sensor fingerprint. (3) Classification of digital videos in an open scenario, that is, where the forensic analyst does not need to have access to the device that recorded the video to find out the origin of the video. The main contribution of the proposed technique eliminates the details of the scene to improve the PRNU fingerprint. It should be noted that these techniques are applied to digital images and not to digital videos. In this work, we show that it is necessary to take this improvement into account to improve the identification of digital videos. Experimental results are also presented that support the validity of the techniques used and show promising results.

Improving multi-view facial expression recognition through two novel texture-based feature representations

Although several automatic computer systems have been proposed to address facial expression recognition problems, the majority of them still fail to cope with some requirements of many practical application scenarios. In this paper, one of the most influential and common issues raised in practical application scenarios when applying automatic facial expression recognition system, head pose variation, is comprehensively explored and investigated. In order to do this, two novel texture feature representations are proposed for implementing multi-view facial expression recognition systems in practical environments. These representations combine the block-based techniques with Local Ternary Pattern-based features, providing a more informative and efficient feature representation of the facial images. In addition, an in-house multi-view facial expression database has been designed and collected to allow us to conduct a detailed research study of the effect of out-of-plane pose angles on the performance of a multi-view facial expression recognition system. Along with the proposed in-house dataset, the proposed system is tested on two well-known facial expression databases, CK+ and BU-3DFE datasets. The obtained results shows that the proposed system outperforms current state-of-the-art 2D facial expression systems in the presence of pose variations.

An efficient copy move forgery detection using adaptive watershed segmentation with AGSO and hybrid feature extraction

Copy-move forgery detection (CMFD) is the process of determining the presence of copied areas in an image. CMFD approaches are mainly classified into two groups: keypoint-based and block-based techniques. In this paper, a new CMFD approach is proposed on the basis of both block and keypoint based approaches. Initially, the forged image is partitioned into non overlapped segments utilizing adaptive watershed segmentation, wherein adaptive H-minima transform is used for extracting the markers. Also, an Adaptive Galactic Swarm Optimization (AGSO) algorithm is used to select optimal gap parameter while selecting the markers for reducing the undesired regional minima, which can increase the segmentation performance. After that, the features from every segment are extracted as segment features (SF) using Hybrid Wavelet Hadamard Transform (HWHT). Then, feature matching is performed using adaptive thresholding. The false matches or outliers can be removed with the help of Random Sample Consensus (RANSAC) algorithm. Finally, the Forgery Region Extraction Algorithm (FREA) is utilized for detecting the copied portion from the host image. Experimental results indicate that the proposed scheme find out image forgery region with Precision = 92.45%; Recall = 93.67% and F1 = 92.75% on MICC-F600 dataset and Precision = 94.52%; Recall = 95.32% and F1 = 93.56% on Bench mark dataset at pixel level. Also, it outperforms the existing approaches when the image undergone certain geometrical transformation and image degradation.

An enhanced block-based Compressed Sensing technique using orthogonal matching pursuit

The theory of compressed sensing asserts that one can recover signals in $$\mathbb {R}^n$$ from far fewer samples or measurements, if the signal has a sparse representation in some orthonormal basis; from non-adaptive linear measurements by solving a $$\mathbb {L}_1$$ norm minimisation problem. The non-adaptive measurements have the character of random linear combinations of the basis or frame elements. However, for large-scale 2D image signals, the randomized sensing matrix consumes enormous computational resources that makes it impractical. The problem has been addressed in the paper as a block compressed sensing (BCS) with sparsity normalization in the transformed domain in the preprocessing stage. The blocks obtained are converted to non-adaptive measurements using identically independent weighted Gaussian random matrices. The feasibility of reconstruction is verified using orthogonal matching pursuit. Simulation results show that better reconstruction performance can be achieved by the proposed technique in comparison with the existing BCS approaches.

An Efficient Two-Level Dictionary-Based Technique for Segmentation and Compression Compound Images

Image data compression algorithms are essential for getting storage space reduction and, perhaps more importantly, to increase their transfer rates, in terms of space-time complexity. Considering that there isn't any encoder that gives good results across all image types and contents, this paper proposed an evolvable lossless statistical block-based technique for segmentation and compression compound or mixed documents that have different content types, such as pictures, graphics, and/or texts.
 
 Derived from the number of detected colors and to achieve better compression ratios, a new well-defined representation of the image is created which nonetheless retains the same image components. With the effort of reducing noise or other variations inside the scanned image, some primary operations are implemented. Thereafter, the proposed algorithm breaks down the compound document image into equal-size-square blocks. Next, inspired by the number of colors detected in each block, these blocks are categorized into a set of six-image objects, called classes, where each one contains a set of closely interrelated pixels that share the same common relevant attributes like color gamut and number, color occurrence, grey level, and others. After that, a new representation of these coherent classes is formed using the Lookup Dictionary Table (LUD), which is the real essence of this proposed algorithm. In order to form distinguishable labeled regions sharing the same attributes, adjacent blocks of similar color features are consolidated together into a single coherent whole entity, called segments or regions. After each region is encoded by one of the most off-the-shelf applicable compression techniques, these regions are eventually fused together into a single data file which then subjects to another compression stage to ensure better compression ratios. After the proposed algorithm has been applied and tested on a database containing 3151 24-bit-RGB-bitmap document images, the empirically-based results prove that the overall algorithm is efficient in the long run and has superior storage space reduction when compared with other existing algorithms. As for the empirical findings, the proposed algorithm has achieved (71.039 %) relative reduction in the data storage space.

Digital image splicing detection technique using optimal threshold based local ternary pattern

Digital images were considered as authentic proof of evidence some years ago but advancement in technology has made image tampering an easy task for every user. Investigation of the digital images for forgery detection, and authenticate their genuineness is need of the hour. To address this issue, the paper proposes a new block-based technique for image splicing detection. In this technique, first the image is converted to YCbCr format and chrominance component of the image is extracted. This component is segmented in overlapping blocks to extract local features. The paper proposes to use a new texture descriptor named as otsu based enhanced local ternary pattern (OELTP) for feature extraction from these blocks. OELTP uses an optimal threshold value to improve the enhanced local ternary pattern (ELTP) texture descriptor, for better detection of image forgery. Further, the paper proposes to use energy for reducing dimensionality of features, instead of using complex computations as used in earlier techniques. Finally, the features are sorted for speedy classification and fed to support vector machine (SVM) for labelling the images either as authentic or forged. The proposed technique has been tested on varying groups of data from the benchmark dataset(s) and has achieved an accuracy upto 98.25%. To demonstrate the superiority of proposed technique, results are also compared with the state-of-the-art techniques.

VLSI Architecture of Block Matching Algorithms for Motion Estimation in High Efficiency Video Coding

High-efficiency video coding (HEVC) is a latest video coding standard and the motion estimation unit is the most important block. The work presents the different types of Matching Criteria for Block-Based Motion Estimation technique in HEVC standard. HEVC requires fast motion estimation algorithms to have better real time performance. The hardware implementation of motion estimation helps to achieve high speed though parallel processing. An improved block matching technique is designed with reduced blocks for HEVC. The proposed method has less execution time where only blocks having motion are compared for prediction computation. The searching time complexity is dependent on the number of blocks that are having motion. The searching time of frame having small motion can be reduced to 80–85% as compared to the traditional full search algorithm. In the paper, sum of absolute difference, mean square error and mean absolute difference are computed to find the best matching algorithm for HEVC. However, SAD has less computational complexity with compare to other matching criteria. The results suggest that proposed motion estimation algorithm has better performance with compare to similar previous work.

A hybrid copy-move image forgery detection technique based on Fourier-Mellin and scale invariant feature transforms

In digital images, the most common forgery is copy-move image forgery in which some region(s) of an image is replicated within the image. The copy-move forgery detection (CMFD) techniques fall under two categories; keypoint-based and block-based. The keypoint-based techniques perform well under rotation and scaling but show very poor performance in the case of smooth images. On the contrary, the block-based techniques perform better in smooth images but are comparatively more time demanding. In this paper, a hybrid technique has been proposed by combining the block-based technique using Fourier-Mellin Transform (FMT) and a keypoint-based technique using Scale Invariant Feature Transform (SIFT). In this technique, the input image to be checked for forgery is first divided into texture and smooth regions. Then the keypoints are extracted from the texture part of the image using the SIFT descriptor, and the FMT is applied on the smooth part of the image. Extracted features are then matched to detect the duplicated regions of the image. The experimental results illustrate that the proposed technique performs better in comparison to other state-of-the-art CMFD techniques under various geometric transformations and post-processing operations in reasonable time.

Multi-Focus Image Fusion and Depth Map Estimation Based on Iterative Region Splitting Techniques.

In this paper, a multi-focus image stack captured by varying positions of the imaging plane is processed to synthesize an all-in-focus (AIF) image and estimate its corresponding depth map. Compared with traditional methods (e.g., pixel- and block-based techniques), our focus-based measures are calculated based on irregularly shaped regions that have been refined or split in an iterative manner, to adapt to different image contents. An initial all-focus image is first computed, which is then segmented to get a region map. Spatial-focal property for each region is then analyzed to determine whether a region should be iteratively split into sub-regions. After iterative splitting, the final region map is used to perform regionally best focusing, based on the Winner-take-all (WTA) strategy, i.e., choosing the best focused pixels from image stack. The depth image can be easily converted from the resulting label image, where the label for each pixel represents the image index from which the pixel with the best focus is chosen. Regions whose focus profiles are not confident in getting a winner of the best focus will resort to spatial propagation from neighboring confident regions. Our experiments show that the adaptive region-splitting algorithm outperforms other state-of-the-art methods or commercial software in synthesis quality (in terms of a well-known Q metric), depth maps (in terms of subjective quality), and processing speed (with a gain of 17.81~40.43%).

Dominating direction based an efficient copy–move image tampering detection technique

ABSTRACTPost-processing operations on a copy–move tampered image makes tampering detection a challenge. Most of the existing techniques are unable to address all the problems simultaneously. This paper presents a block-based technique, which addresses all these problems effectively. Here, at first, the approximate band coefficients are extracted using Contourlet Transform for further steps as it reduces the dimensionality as well as the effects of noise. Then Harris’ matrix is used to find the dominant direction of the block to counter the rotation. Finally, singular values of blocks are calculated using singular value decomposition, which works as a feature vector. The efficacy of the proposed algorithm has been verified through extensive simulations and comparisons with several other challenging copy–move image tampering detection techniques.

Efficient data retrieval using adaptive clustered indexing for continuous queries over streaming data

The Modern era has highly dynamic, heterogeneous and massive data volumes, generated from sensor networks, social media and telecommunications, stock market analyses and the Internet, etc. makes constant query processing quite challenging in processing real-time data, which exist as streams and undergo dynamic changes. Large volumes of data can be efficiently handled by partitioning them into clusters followed by Indexing. An efficient clustering and indexing method is required to process continuous queries for retrieving data streams. A new index structure called adaptive clustering and block-based indexing (ACBBI) is proposed, which is a fusion of cluster-based and block-based techniques to process continuous queries. The incoming data are clustered and stored as blocks using the adaptive clustering method and further indexed by the adaptive indexing approach. Livestock market values that are time variant are used for experimentation. The experimental analysis demonstrates that the ACBBI tree structure significantly decreases half of the space cost, scales better with increasing data size and improves the retrieval rate 30% more than an existing CKDB approach.

Feature ranking and selection used in a machine learning framework for predicting uniformity of printed pages

In our previous work [1,2], we presented a block-based technique to analyze printed page uniformity both visually and metrically. In this paper, we introduce a new sets of tools for feature ranking and selection. The features learned from the models are then employed in a Support Vector Machine (SVM) framework to classify the pages into one of the two categories of acceptable and unacceptable quality. We utilize three methods in feature ranking including F-score, Linear-SVM weight, and Forward Search. The first two methods are filter methods while the last is categorized as a wrapper approach. We use the result from the wrapper method and information from the filter methods as confidence scores in our feature selection framework

An adaptive block-based matching algorithm for crowd motion sequences

For crowd analytics and surveillance systems, motion estimation is an essential first step. Lots of crowd motion estimation algorithms have been presented in the last years comprising pedestrian motion. However, algorithms based on optical flow and background subtraction have numerous limitations such as the complexity of the computation in the presence of high dense crowd and sudden motion changes. Therefore, a novel estimation algorithm is proposed to measure the motion of crowd with less computational complexity and satisfy the real time requirements. The proposed algorithm is based on block-based matching, particle advection, and social force model. By the block-based matching, the motion is estimated in each frame, and the corresponding motion field is created. The particle advection process provides more information about the behavior of pedestrians groups, their tracked trajectories and the boundary of each group segment. Relying on the social force model, a predicted direction of the motion vectors (MV) could be measured significantly. Subsequently, the block-based technique is combined with the social force model to obtain the accurate motion vector with the less possible number of search points. The experimental results indicate that the proposed method achieves high performance by reducing the search points, particularly when many collision situations or obstacles exist in the scenes. Considering the reduction in the computational complexity, the quality of degradation is very low. In all cases, average PSNR degradation of the proposed algorithm is only 0.09.

Non-Overlapping Block-based Parametric Forgery Detection Model

Modification of a digital image by adding or removing some of its elements using a wide variety of image processing tools results in image forgery. As a result authentication of originality of a digital image is becoming a challenging task. Copy-paste forgery is one of the forgeries belonging to context based forgery. Copy-Paste Forgery Detection (CPFD) aims at finding regions that have been copied and pasted within the same or different image. A small change in the image may change statistical parameters that can be analysed for initial assessment of the forgery. In the present research study, a parametric forgery detection model using nonoverlapping block-based technique is developed to ascertain the copy-paste forgery in a given digital image. Statistical parameters of the input image are computed, analysed and compared with those of the forged image. The results show that the proposed model identifies the forged area of the given image and works well with low to moderate copy-paste forgery. The results obtained can be used as the initial verification of the images for forgery and to enhance the forgery detection process by identifying most likely cases of possible image forgeries. The proposed model is tested with large domain of images having different dimensions and for detecting forgery within an image. However, the model has limitations with certain geometrical transformations. General Terms Digital Image Forgery Detection Techniques.

Unconstrained face verification with a dual-layer block-based metric learning

In this paper, a dual-layer block-based metric learning technique is proposed to better discriminate the face image pairs and accelerate the overall verification process under the unconstrained environment. The input images are processed as blocks to provide a richer base of face features. Our proposed method is formed by two layers, in which the first layer assists in extracting the compact block-based descriptors without the existence of full class label information and to refine the within-class and between-class scatter matrices while the second layer integrates the face descriptors of all blocks. The proposed scheme has computational advantage over the single metric learning method while it exploits the correlations among the multiple metrics from different descriptors. The performance of our proposed method is evaluated on the Labeled Faces in the Wild database and achieves an improved performance when compared with the state-of-the-art methods in terms of verification rate and computation time.

Multi-focus image fusion using best-so-far ABC strategies

Multi-focus image fusion is a process of combining a set of images that have been captured from the same scene but with different focuses in order to construct an additional sharper image. This process plays an important role in the image processing and machine vision fields. Various algorithms have been developed for this task. The key challenge in the design of multi-focus image fusion algorithms is how to evaluate the local content information of each image from the source images. Simple, but effective, block-based techniques at pixel level are widely used for multi-focus image fusion. However, a fixed block size may not be applicable to every application. A block size that is too small or too large is also not desirable. Hence, optimization of the block size is necessary in order to obtain a fused image that comprises the sharper parts of the source images. Recently, a number of techniques based on evolutionary computation have been applied to block-based multi-focus image fusion. The artificial bee colony (ABC) algorithm is one of the more popular evolutionary computational approaches used to find an optimal solution. In this paper, an efficient and robust block-based multi-focus image fusion method based on the optimal selection of sharper image blocks from source images using best-so-far ABC strategies is proposed. Experiment results show that the proposed method is able to provide good results and outperforms other conventional methods, both visually and quantitatively.

Block-Based Compressed Sensing for Neutron Radiation Image Using WDFB

An ideal compression method for neutron radiation image should have high compression ratio while keeping more details of the original image. Compressed sensing (CS), which can break through the restrictions of sampling theorem, is likely to offer an efficient compression scheme for the neutron radiation image. Combining wavelet transform with directional filter banks, a novel nonredundant multiscale geometry analysis transform named Wavelet Directional Filter Banks (WDFB) is constructed and applied to represent neutron radiation image sparsely. Then, the block-based CS technique is introduced and a high performance CS scheme for neutron radiation image is proposed. By performing two-step iterative shrinkage algorithm the problem of L1 norm minimization is solved to reconstruct neutron radiation image from random measurements. The experiment results demonstrate that the scheme not only improves the quality of reconstructed image obviously but also retains more details of original image.

English

2 ABSTRACT: Video compression uses Motion Estimation (ME) technique to achieve very high compression ratio. However ME is the most computational part of video encoding. Therefore development of many fast ME algorithms have been done to reduce the computations. Block-based techniques are the most popular one due to its simplicity. This research paper implements five Block Matching Algorithms (BMAs) namely Exhaustive Search (ES), Three Step Search (TSS), New Three Step Search (NTSS), Four Step Search (4SS) and Diamond Search (DS).Their performances are compared and analyzed based on the number of computations and PSNR.