Non-edge Pixels Research Articles

Increasing embedding capacity of stego images by exploiting edge pixels in prediction error space

In the field of data concealing, edge detection techniques are frequently employed, particularly for improving image quality and data security. These methods, however, have a lower embedding capacity. In order to take advantage of more edge pixels, many strategies are used nowadays. These schemes either combine the output from multiple edge detectors or enlarge the edges of an edge image by dilating. Even so, if the amount of data is vast, the techniques might not be able to conceal all of it. Therefore, a novel strategy for edge exploitation is still needed to regulate the effectiveness of edge detection-based data-hiding strategies. By using edge detectors in the prediction error space, we utilized more edge pixels in this study (PES). Applying a predictor on the cover image and then calculating the prediction errors, we prepared the PES. The edges in PES were then marked using the edge detector. The edge-error corresponding pixels received more information than the relevant pixels that did not create an edge-error. Additionally, we combined the results from different edge detectors to produce more edges, which does help to achieve a higher embedding capacity. We implanted x number of secret bits in edge pixels and y number of bits in non-edge pixels where x>y. The simulation results show that the proposed scheme outperforms its rivals on all performance-measuring criteria, including payload, stego image quality, and resistance to attack.

An Image Edge Detection Algorithm Based on an Artificial Plant Community

Image edge detection is a difficult task, because it requires the accurate removal of irrelevant pixels, while retaining important pixels that describe the image’s structural properties. Here, an artificial plant community algorithm is proposed to aid in the solving of the image edge detection problem. First, the image edge detection problem is modeled as an objective function of an artificial plant community searching for water sources and nutrients. After many iterations, the artificial plant community is concentrated in habitable areas that are rich in water sources and nutrients, that is, the image edges, and the nonhabitable zones that are not suitable for living are deserted, that is, the nonedges. Second, an artificial plant community algorithm is designed to solve the objective function by simulating the growth process of a true plant community. The living behavior of the artificial plant community includes three operations: seeding, growing, and fruiting. The individuals in the plant community also correspond to three forms, namely seeds, individuals, and fruit. There are three fitness comparisons in each iteration. The first fitness comparison of each iteration is carried out during the seeding operation. Only the fruit with higher fitness levels in the last iteration can become seeds, while the fruit with low fitness levels die, and some new seeds are randomly generated. The second fitness comparison is implemented in the growing operation. Only the seeds with higher fitness levels can become individuals, but the seeds with lower fitness levels will die; thus, the community size will decrease. The third fitness comparison is in the fruiting operation, where the individual with the greatest fitness can produce an identical fruit through parthenogenesis, and the individuals with higher fitness levels can learn from each other and produce more fruit, so the population size can be restored. Through the continuous cycle of these three operations, the artificial plant community will finally determine the edge pixels and delete the nonedge pixels. Third, the experiment results reveal how the proposed algorithm generates the edge image, and the comparative results demonstrate that the proposed artificial plant community algorithm can effectively solve the image edge detection problems. Finally, this study and some limitations are summarized, and future directions are suggested. The proposed algorithm is expected to act as a new research tool for solving various complex problems.

A dual-scale morphological filtering method for composite damage identification using FBP

Delamination is the main type of defect that reduces the performance of carbon fiber-reinforced polymer (CFRP) composite materials and can be detected by traditional linear ultrasonic methods. However, the traditional ultrasonic inspection method only analyzes its position through the waveform change, and it is difficult to intuitively and accurately study the specific shape and position of the defect in the form of images. This study obtains delamination defect images of CFRP composite plates through fan-beam ultrasonic computed tomography (UCT) and finite element simulation. The Filtered back-projection method based on UCT is used as the basic algorithm for reconstructing layered defect images of CFRP plates. To obtain accurate defect size and location, a novel dual-scale morphological filtering (DSMF) method combined with Canny operator is proposed for edge detection of delamination defects. The DSMF method alternately filters by combining two structuring elements of different sizes while suppressing noise as much as possible and preserving the edge details of the image. As the results verify, the use of the DSMF method combined with Canny edge-detection for edge detection significantly reduces the average recognition error from 16.83% to 4.37% compared to existing methods. The results show that the DSMF method can effectively reduce the influence of artifact interference and non-edge pixels on defect-recognition; thus, the size and contour information of the obtained defects are more accurate than existing methods. Experimental studies using realistic composite plates further validated the proposed method. The method of ultrasonic tomography combined with Canny edge detection based on a DSMF method can be well applied to edge detection and contour recognition of layered defects in CFRP materials.

Image Interpolation with Regional Gradient Estimation

This paper proposes an image interpolation method with regional gradient estimation (GEI) to solve the problem of the nonlinear interpolation method not sufficiently considering non-edge pixels. First, the approach presented in this paper expanded on the edge diffusion idea used in CGI and proposed a regional gradient estimation strategy to improve the problem of gradient calculation in the CGI method. Next, the gradient value was used to determine whether a pixel was an edge pixel. Then, a 1D directional filter was employed to process edge pixels while interpolating non-edge pixels using a 2D directionless filter. Finally, we experimented with various representative interpolation methods for grayscale and color images, including the one presented in this paper, and compared them in terms of subjective results, objective criteria, and computational complexity. The experimental results showed that GEI performed better than the other methods in an experiment concerning the visual effect, objective criteria, and computational complexity.

Payload and quality augmentation using steganographic optimization technique based on edge detection

Information security is one of the most significant processes that must be taken into account when confidentially transferring information. This paper introduces a steganography technique using the edge detection method. It focused on three basic and important aspects’ payload, quality and security. Well-known edge detectors were used to generate as many edge pixels as possible to hide data and achieve the highest payload. The least significant bit (LSB) algorithm has been improved by extending the bits used to embed between 2-4 bits in smooth and sharp areas. To increase security, the transaction between the two parties is based on dividing the key and the cover image into several parts and agreeing on the type of edge detection.The experiments achieved the maximum load, for instance with a fuzzy edge detector, at first, embedding in 4 bitplanes if edge pixel and in 2 bitplanes if non-edge pixel, the peak signal-to-noise ratio (PSNR) increased from 43.580to 45.790. At second, embedding in 2 bitplanes if edge pixel, and in 4 bitplanes if non-edge pixel, the PSNR decreased between 38.433-41.593. The suggested scheme achieved a high pay load to embed in the cover image and according to human perception, it preserved the nature of the original image.

Multi-residual Connection Network for Edge Detection

CNN-based methods have improved the performance of edge detection in recent years. However, the edge map predicted by the neural network has a problem of thickness. We analyze that this is due to the training problem caused by the unbalanced distribution of edge and non-edge pixels. We propose to use residual connections to solve the class-imbalanced training problem in edge samples. Moreover, we introduce a new biased cross-entropy loss function to better train the edge detection network, which will adjust the weights according to the ratio of edges and non-edges pixels. Compared to other methods, our method predicts clearer and crisp edges. We set up multiple long and short residual connections in the network to establish various information propagation pathways. This makes the final prediction of our method not only with contour information but also have rich details. We evaluated our method on the BSDS500 and NYUD datasets and showed promising results.

Mini Kirsch Edge Detection and Its Sharpening Effect

In computer vision, edge detection is a crucial step in identifying the objects’ boundaries in an image. The existing edge detection methods function in either spatial domain or frequency domain, fail to outline the high continuity boundaries of the objects. In this work, we modified four-directional mini Kirsch edge detection kernels which enable full directional edge detection. We also introduced the novel involvement of the proposed method in image sharpening by adding the resulting edge map onto the original input image to enhance the edge details in the image. From the edge detection performance tests, our proposed method acquired the highest true edge pixels and true non-edge pixels detection, yielding the highest accuracy among all the comparing methods. Moreover, the sharpening effect offered by our proposed framework could achieve a more favorable visual appearance with a competitive score of peak signal-to-noise ratio and structural similarity index value compared to the most widely used unsharp masking and Laplacian of Gaussian sharpening methods. The edges of the sharpened image are further enhanced could potentially contribute to better boundary tracking and higher segmentation accuracy.

Mini Kirsch Edge Detection and Its Sharpening Effect

In computer vision, edge detection is a crucial step in identifying the objects’ boundaries in an image. The existing edge detection methods function in either spatial domain or frequency domain, fail to outline the high continuity boundaries of the objects. In this work, we modified four-directional mini Kirsch edge detection kernels which enable full directional edge detection. We also introduced the novel involvement of the proposed method in image sharpening by adding the resulting edge map onto the original input image to enhance the edge details in the image. From the edge detection performance tests, our proposed method acquired the highest true edge pixels and true non-edge pixels detection, yielding the highest accuracy among all the comparing methods. Moreover, the sharpening effect offered by our proposed framework could achieve a more favorable visual appearance with a competitive score of peak signal-to-noise ratio and structural similarity index value compared to the most widely used unsharp masking and Laplacian of Gaussian sharpening methods. The edges of the sharpened image are further enhanced could potentially contribute to better boundary tracking and higher segmentation accuracy.

Advanced image steganography based on exploiting modification direction and neutrosophic set

A new variant of image steganography based on exploiting modification directions (EMD) named advanced EMD (AEMD) is introduced. In this method the secret digits in mn -ary notional systems are embedded into a group of n pixels of the cover image. To increase data hiding capacity, edge masking characteristics of human visual system is exploited to embed more bits at image edge pixels than non-edge pixels. To do this, a neutrosophic set edge detector, named as MNSED is also introduced to classify cover image into 2 × 2 non-overlapping edge and non-edge blocks without any overhead information. Hence the secret digits in two different bases are embedded into the edge and non-edge blocks. Experimental results show that the proposed method provides higher embedding capacity and better quality compared to the existing schemes, while its resistance to steganographic attack is still higher.

AGRNet: Adaptive Graph Representation Learning and Reasoning for Face Parsing.

Face parsing infers a pixel-wise label to each facial component, which has drawn much attention recently. Previous methods have shown their success in face parsing, which however overlook the correlation among facial components. As a matter of fact, the component-wise relationship is a critical clue in discriminating ambiguous pixels in facial area. To address this issue, we propose adaptive graph representation learning and reasoning over facial components, aiming to learn representative vertices that describe each component, exploit the component-wise relationship and thereby produce accurate parsing results against ambiguity. In particular, we devise an adaptive and differentiable graph abstraction method to represent the components on a graph via pixel-to-vertex projection under the initial condition of a predicted parsing map, where pixel features within a certain facial region are aggregated onto a vertex. Further, we explicitly incorporate the image edge as a prior in the model, which helps to discriminate edge and non-edge pixels during the projection, thus leading to refined parsing results along the edges. Then, our model learns and reasons over the relations among components by propagating information across vertices on the graph. Finally, the refined vertex features are projected back to pixel grids for the prediction of the final parsing map. To train our model, we propose a discriminative loss to penalize small distances between vertices in the feature space, which leads to distinct vertices with strong semantics. Experimental results show the superior performance of the proposed model on multiple face parsing datasets, along with the validation on the human parsing task to demonstrate the generalizability of our model.

Image steganography based on Kirsch edge detection

Conventional steganography methods fabricate the secret information into the cover pixels without analyzing the pixel intensities of an image. As a result, some minor pixel level manipulations may lead to huge visual distortion in the stego-image. To this end, in this paper, a novel steganographic scheme based on Kirsch edge detector is proposed. The aim of the scheme is to maximize the payload by embedding more secret bits into edge pixels and fewer bits into the non-edge pixels. The proposed scheme has three major phases: construction of edge image, embedding and extraction. The first phase deals with the construction of masked image from the cover image, and in turn, edge image from the masked one. The second phase deals with the decomposition of the cover image into a set of triplet of pixels and then embedding of (x + y + 1) bits of secret data into each triplet of pixels to obtain the stego-image. Here, ‘x’ and ‘y’ are not fixed as the edge information of each triplet changes incessantly. The third or last phase deals with the extraction of the secret information from the stego-image using the reverse process. Simulation results on some standard images ensure that the proposed method achieves higher payload and better image quality compared to the conventional steganographic schemes. Furthermore, the Kirsch edge detector is able to produce more number of edge pixels compared to the traditional edge detectors; and, hence the proposed scheme also outperforms the existing edge-based methods in terms of payload.

Developing a feature decoder network with low-to-high hierarchies to improve edge detection

Low-to-high hierarchical convolutional features can significantly improve edge detection. This paper proposes a feature decoder-based algorithm that employs a Feature Decoder Network (FDN) to extract more information within limited Convolutional Neural Network (CNN) features. Previous studies applied convolutional elements by weight fusion, but we measure a feature decoder as a pyramid by qualifying convolutional layers. The feature decoder fuses CNN features of adjacent layers to judge the edge and non-edge pixels, which can learn the relationship and distinction between low-level edge hierarchies and high-level semantic hierarchies. Furthermore, we use Gaussian blur labels to train the network to optimize network convergence and training. From the experimental results, our proposed algorithm performs better on the BSDS500 (average accuracy (AP) of 0.865) and NYUD (OIS F-measure of 0.775) datasets compared to the state-of-the-art algorithms, including RCF.

A novel minimal distortion-based edge adaptive image steganography scheme using local complexity

The advantage of spatial domain image steganography techniques is their capacity to embed high payloads of data by directly modifying image pixels. While these techniques have a high-embedding capacity, they often create visual and statistical distortion in smoother regions. Most existing edge steganography techniques divide an image into blocks and insert data by processing the blocks in a linear order, but these method also has multiple drawbacks. First, if the selected block has an insufficient number of edge pixels, it may result in multiple blocks being processed. Second, at high embedding rates, the method creates severe distortion as multiple message bits are hidden in edge pixels and surrounding non-edge pixels without analyzing the statistical dependencies and correlation of pixels, compromising data security. The aim of the proposed method is to construct a Block-wise Edge Adaptive Steganography Scheme (BEASS) using textured regions, particularly edges and surrounding pixels. This scheme dynamically chooses the region to embed messages using a local complexity measure of Standard Deviation. It offers high payload, minimal distortion embedding by hiding three message bits into edge pixels using the minimal Mean Square Error to determine the embedding capacity of neighboring non-edge pixels within the block to preserve the statistical dependencies. The practical merit of this approach was validated and compared with existing algorithms, and experimental results find that the proposed method surpasses IQM tests, achieves a high PSNR of61 $\sim $ 65, proves to be robust against kurtosis and skewness distortion, resists histogram attack, RS steganalysis and high dimensional ensemble classifier at 80% block modifications.

RETRACTED ARTICLE: A stable image steganography: a novel approach based on modified RSA algorithm and 2–4 least significant bit (LSB) technique

Nowadays, due to the rapid technological growth of information technology billions of data transferred through the network around the world in a fraction of seconds. To protect the transferred information from unauthorized access, security mechanism is required. At this point, significant role is performed by a technology called information security. So, to provide security, various techniques have been evolved such as cryptography, steganography and digital watermarking. Cryptography converts the message into scrambled format whereas in steganography the secret message is buried in the digital media. In this paper, both cryptography and steganography is combined to safeguard the information. First, the message is encrypted with the help of revised RSA algorithm to convert the message into secret message. Second, the cover image is segmented into edge and non-edge pixels using Canny edge detection technique and then N1(two) and N2 (four) bits of the secret messages are buried into the edge and non-edge pixel areas using LSB technique. Finally, the essential parameters like N1, N2 and the length values are implanted into the last four pixels of the cover-image to obtain the stego-image. At the recipient side, the hidden data are extracted from the stego-image after extracting N1, N2 and length information. Then, the extracted secret data is decrypted using revised RSA algorithm. Modified RSA algorithm improves the security by using two pairs of keys. At last, the various performance metrics parameters such as entropy, MSE, peak signal to noise ratio and histogram are measured. It is observed that the proposed system attains improved efficiency, security and imperceptibility properties of information hiding from the results.

A new fuzzy rule based pixel organization scheme for optimal edge detection and impulse noise removal

Fuzzy sets provide a framework for incorporating human knowledge as an efficient unsupervised machine learning tool for problem solving. The approach discussed in this paper introduces a generalized transfer learning scheme using rule based fuzzy logic for edge detection in digital images. The spatial domain statistical properties of the image are explored as training data set and expressed in fuzzy format to obtain a decision function for optimal edge detection along with reduction of impulse noise. During fuzzy inference process, a specific linguistic value in input fuzzy set is selected in order to obtain an optimal range of second order difference which discriminates the edge pixels from the non-edge pixels. The proposed fuzzy rule based optimal edge pixel detection method in the presence of random valued impulse noise tends to sufficiently extract the edge pixels with out boosting the noisy pixels. The effectiveness of the proposed fuzzy rule based edge detection scheme is verified by testing it on various standard test images and comparing with existing edge detection techniques at different noise densities.

Edge Detection using Guided Image Filtering and Enhanced Ant Colony Optimization

Abstract Edge detection is an important process in many classes of engineering problems. An edge is a boundary between two regions. In the edge detection process, the input is a grayscale image while the output image is a binary image showing edge and non-edge pixels. As a state of art, methods are based on gradient operator, but the accuracy level is poor. Therefore, in recent past soft computing based methods are proposed which are more accurate, still, these methods fail to detect some of the true edges. In this paper, the two-level approach is adopted for edge detection, firstly image edges are enhanced using guided image filtering, and secondly on these enhanced images enhanced ant colony optimization method is applied for edge detection.

Noise-robust color edge detection using anisotropic morphological directional derivative matrix

In this paper, a color edge detector using the anisotropic morphological directional derivatives (AMDDs) is presented to detect edges in color images corrupted by Gaussian or impulsive noise. The AMDD matrix, robust to impulsive noise owing to the underlying morphological operators, is constructed to represent edge information at each pixel of a color image. The color edge strength map and color edge direction map of a color image are extracted by spatial and directional matched filtering and singular value decomposition of the AMDD matrices. Embedding them in the route of the Canny detector yields a noise-robust color edge detector. Moreover, a color image database with groundtruths (GTs) of edges are built. The GT of a color image is generated in three steps. First, the contours and results of multiple color edge detectors are fused into a candidate edge map (CEM). Next, the CEM, the original image, and a special software for edge modification are sent to twenty experienced observers to modify the CEM. Finally, their results are used to yield the edge pixels, non-edge pixels, and don't care regions in the GT by the voting rule. The proposed detector is compared with existing color edge detectors on the database.

Modified Bird swarm algorithm for edge detection in noisy images using fuzzy reasoning

ABSTRACTA new bio-inspired edge detection approach is proposed to deal with the noisy images using a combination of bird swarm algorithm (BSA) and fuzzy reasoning. BSA is based on the behaviour of the birds. The birds fly through each pixel while they forage for the food and detect the edge pixels and noisy pixels that fall in their path. The direction in which the birds fly is found using fuzzy rule-based system. The pixels are classified as edge and non-edge pixels using the concept of thresholding. The noisy pixels are removed by the birds using fuzzy impulse noise detection and reduction method. The technique has been evaluated on two standard image datasets for quantitative and qualitative analysis. The results clearly indicate significant improvement over the other bio-inspired approaches, namely, Genetic Algorithm (GA), Particle Swarm Optimisation (PSO), Ant Colony Optimisation (ACO), Bacterial Foraging Algorithm (BFO), Deep Learning approach, fuzzy with BFO for noisy images, Neuro-fuzzy approach and PSO particularly for images having impulse noise in terms of Entropy, Kappa Value, Pratt’s Figure of Merit (FoM) and Structural Similarity Index Measure (SSIM). The proposed approach works well to detect the continuous, thin and smooth edges in the presence of 5–40% noise density.

High payload image steganography based on Laplacian of Gaussian (LoG) edge detector

Abstract Image Steganography is the method of concealing secret information into the digital image, and LSB replacement policy is recognized as the foremost and widely used approach. In traditional LSB based scheme, the secret bits are fabricated into the LSB position of each pixel without making a prior analysis of the image contents. As a consequence, the visual quality as well as the security of the stego-image becomes a concern even to achieve a low embedding rate. In this paper, the embedding of secret data has been made by classifying the pixels of the cover image into following two categories: edge and non-edge. This classification is made by applying the Laplacian of Gaussian (LoG) edge detector over the gray-scale images. It has been observed that the non-edge pixels get affected severely compared to edge pixels at higher embedding rate and therefore, more bits are embedded in edge pixels rather than in the non-edge pixels. The disadvantage of existing edge detection based Steganography schemes is the embedding of (extra) edge information besides the actual embedding. Unlikely, the proposed scheme proficiently avoids the burden of embedding of extra edge information and utilizes the embedding space fully. Experimental results ensure that the proposed scheme achieves higher payload and better stego-image quality compared to existing schemes.

A novel approach for color image segmentation based on region growing

In this paper, a novel approach for obtaining all possible uniform regions in the color image is proposed. The proposed approach integrates a color edge detection method; image partitioning; the initial seeds and thresholding region growing; the average overlap metric (AOM) and voting algorithms. It starts by decomposing the color image into less complicated component images. The edges of the color image are detected to extract the non-edge pixels during the region growing processes. Then, a source image is partitioned into cells while seeds are obtained by applying the local search algorithm in the image histogram. The growing processes are improved by color image thresholding algorithm which is necessary for finding the homogeneity criterion to merge similar pixels. The seeds and the homogeneity criterion values are the input to the region growing method to segment an image into regions; some of them are overlapped or been redundant. The AOM algorithm is applied to classify the redundant regions based on pixel similarity. These regions are fed to voting technique in order to produce region of points whose have similar values to utilize the compactness of the clusters forming these uniform regions. Experimental results are conducted using different color images with different sizes. Moreover, the proposed method is experimented by different noisy images and is compared with the well-known existing methods to prove its efficiency. The obtained results reveal the accuracy and stability of proposed technique and its superiority over other three well-known existing methods.