Tetrolet Transform Research Articles

Underwater Fused Image Classification Using Deep Learning Based Resnet and Hybrid PSO + HHO Model

In image fusion, multiple images are combined into one with minimal distortion and data loss. Image fusion forms a highly configured image with good results. The image fusion technique is employed in various domains, including remote sensing, robotics, medical applications, and underwater image processing. The focus of this study is on a unique underwater image fusion approach that allows for greater flexibility in the construction of fusion criteria. The characteristics of the input images are extracted using a Modified Tetrolet Transform (MMT), which may be employed together or independently. Our aim is to apply image deep learning algorithms such as GoogleNet, AlexNet, and ResNet for fusion to the original image and optimized algorithm PSO and HHO to be used for optimizing the fused image. Finally, the images are classified as good quality images and poor-quality images. The ResNet with hybrid HHO and PSO method has high efficiency in image fusion, according to numerical data the proposed model with good quality attains accuracy, sensitivity, specificity, precision, and F1-measure are 96.32, 94.25, 95.73, 98.34, and 95.55. in addition, the poor-quality images attain accuracy, sensitivity, specificity, precision, and F1-measure are 96.22, 94.15, 95.63, 98.24, and 95.45 This method optimizes the exposure of the dark areas, increases contrast, and preserves and enhances the edges. Our image trial findings show that this technique substantially improves the underwater image quality.

A Robust Illumination and Intensity invariant Face Recognition System

Face recognition has achieved more attention in computer vision with the focus on modelling the expression variations of human. However, in computer vision system, face recognition is a challenging task, due to variation in expressions, poses, and lighting conditions. This paper proposes a facial recognition technique based on 2D Hybrid Markov Model (2D HMM), Cat Swam Optimization (CSO), Local Directional Pattern (LDP), and Tetrolet Transform. Skin segmentation method is used for pre-processing followed by filtering to extract the region of interest. Resultant image is fed to proposed feature extraction method comprising of Tetrolet Transform and LDP. Extracted features are classified using proposed classifier “CSO trained 2D-HMM classification method”. To prove the superiority of method, four face datasets are used, and comparative results are presented. Quantitively results are measured by False Acceptance Rate (FAR), False Rejection Rate (FRR) and Accuracy and the values are 0.0025, 0.0035 and 99.65% respectively

ScatT-LOOP: scattering tetrolet-LOOP descriptor and optimized NN for iris recognition at-a-distance.

Iris Recognition at-a Distance (IAAD) is a major challenge for researchers due to the defects associated with the visual imaging and poor image quality in dynamic environments, which imposed bad impacts on the accuracy of recognition. Thus, in order to enable the effective IAAD, this paper proposes a new method, named, Chronological Monarch Butterfly Optimization (Chronological MBO)-enabled Neural Network (NN). The recognition of iris using NN is trained with the proposed Chronological MBO, which is developed through the combination of Chronological theory in Monarch Butterfly Optimization (MBO). The recognition becomes effective with the automatic segmentation and the normalization of iris image on the basis of Hough Transform (HT) and Daugman's rubber sheet model followed with the process of feature extraction with the developed ScatT-LOOP descriptor, which is the integration of scattering transform (ST), Local Optimal Oriented Pattern (LOOP) descriptor, and Tetrolet transform (TT). The developed ScatT-LOOP descriptor extracts the texture as well as the orientation details of image for effective recognition. The analysis is evaluated with the CASIA Iris dataset with respect to the evaluation metrics, accuracy, False Acceptance Rate (FAR), and False Rejection Rate (FRR). The proposed method has the accuracy, FRR, and FAR of 0.97, 0.005, and 0.005, respectively. The experimental results proved that the proposed method is effective than the existing methods of iris recognition.

Research on Infrared and Visible Image Fusion Based on Tetrolet Transform and Convolution Sparse Representation

Image fusion is a visual enhancement technique that combines source images from different sensors to produce a more robust and informative fused image for subsequent processing or decision making. Infrared and visible light images share complementary properties that enable the production of robust and informative fused images. This paper proposed an infrared and visible image fusion method that improved the tetrolet framework to improve infrared and visible image fusion quality. First, the source image is enhanced by bicubic interpolation. The improved tetrolet transform then decomposes the enhanced source image; the high-frequency components are fused by convolutional sparse representation theory and combined with corresponding rules, and the low-frequency components are fused by defining ISER descriptors. Finally, we use the inverse transform to reconstruct the fused image. Qualitative and quantitative experimental results on five groups of typical infrared and visible image datasets demonstrate the proposed method's effectiveness. The proposed method exhibits better performances on subjective vision and objective indexes compared with the other state-of-the-art methods.

Edge preserved multispectral image compression using PCA and hybrid transform

In multispectral image compression, it is difficult to obtain sparse approximation for images with rich details. The existing methods produce better results under some constraints on image content. In order to obtain robust multispectral image compression, the Principal Component Analysis (PCA) method is used to reduce the spectral redundancy and the sparse approximation is obtained by using Extended Shearlet Transform (EST) and Tetrolet Transform(TT). The anisotropic property of EST is used to preserve smooth images with global structures of an image whereas the TT is used to preserve the local structures. The performance of proposed method is compared with the existing methods in terms of rate distortion and information preservation perspectives. The Compression Ratio (CR) and Peak Signal to Noise Ratio (PSNR) are used as rate-distortion measure. The Mean Structural Similarity Index Metric (MSSIM) and Kappa coefficient (K) are information preservation measures. The simulation results show that the proposed EPMI-HT method outperforms the existing hybrid methods for all kinds of image content at high CR with retaining edge information.

RETRACTED ARTICLE: Multilevel Tetrolet transform based breast cancer classifier and diagnosis system for healthcare applications

Breast cancer is a deadly disease and nowadays affects middle-aged women. Doctors use various methods to find out the presence of micro calcifications. In general, images of digital mammogram are used for cancer diagnosis. But the precision that determines the presence of cancer requires multiple restrictions. An improved medical decision support system for the classification of mammogram images using Unsupervised Test Vector Optimization (UTVO) based classifier is developed for health care and is presented in this work. The median filtering in the preprocessing stage removes the speckle, salt and pepper noises for mammogram image enhancement and the multilevel Tetrolet transformation based Matrix vectors are obtained for feature extraction. Multilevel Tetrolet transformation based features (MTF) are used for the selection of texture features. The UTVO classification approach proposed for identifying breast cancer exploits the potential and analyzes its performance after classification of benign and malignant cases. Finally, the UTVO classification system’s memory and computational time has been used to weigh the optimal weight and used to analyze the performance without compromise. The proposed system is expected to provide the best support for doctors to achieve maximum accuracy and improvement in health care.

Fusion Algorithm for Near-Infrared and Color Visible Images Based on Tetrolet Transform

To address the problems of low contrast,loss of detail,and color distortion after near-infrared and color visible image fusion,an algorithm for near-infrared and color visible image fusion based on the Tetrolet transform and pulse coupled neural network(PCNN)is proposed.First,the color visible light source image is transformed into a hue saturation intensity space,where each component is relatively independent,and its brightness component is decomposed into a near-infrared image by Tetrolet decomposition.Subsequently,a fusion rule for expectation maximization likelihood estimation of the potential distribution from a given incomplete dataset is proposed.A self-adaptive PCNN model with a Sobel operator that automatically adjusts the threshold is used as a fusion rule,and the processed high-frequency and low-frequency images are fused by Tetrolet inverse transformation as brightness images.An adaptive stretching method for the saturation component is proposed to solve the problem of image saturation decline.The processed components are mapped back to red–green–blue space to complete the fusion.The proposed algorithm was compared with several efficient fusion algorithms.The experimental results show that the image obtained by this method has clear details and improved color contrast.It has obvious advantages in image saturation,color restoration performance,structural similarity,and contrast.

RETRACTED ARTICLE: Lossless medical image compression algorithm using tetrolet transformation

In many analytical applications the reuse of medical images plays an important role. In order to store the images, image compression is the technique used to save the storage space and the storage time. The image compression is used in image transmission for the fast transmission. Also in health care sector the compressed image helps to make it complete computerized. The noise interference is the major problem in the image compression. Thus some compression techniques are used to remove the noise added in medical images. This will improve the quality of the compressed image. In image compression the medical image can be processed in deep degree by de-noising, edge preservation and high compression rate. The aim of the paper is develop an efficient computational tetrolet transform, which can be used in lossless compression. For analysis purpose various wavelet techniques are used and result is compared with this techniques. From the analysis it is inferred that by proposed technique is able to achieve a high quality images with low noise.

Fusion of Underwater Images based on Tetrolet Transform &Color Correction Algorithms

Fusion of Underwater Images based on Tetrolet Transform &Color Correction Algorithms

Image Retrieval Scheme Using Quantized Bins of Color Image Components and Adaptive Tetrolet Transform

In this paper, a three stage hierarchical image retrieval scheme using a color, texture and shape visual contents (or descriptors) is proposed, since single visual content is not produce an adequate retrieval results effectively. This scheme has reduced the searching space during the image retrieval process at a certain extent due to the hierarchical mode. In initial stage, the shape feature descriptor has been computed by simple fusion of histograms of gradients and invariant moments of segmented image planes. The shape based retrieval process has reduced the search space by discarding the non-relevant images from the universal dataset (or original dataset) effectively and kept the retrieved images into the intermediate dataset. In the second stage, the texture feature descriptors have been computed from the intermediate sub-image dataset by applying the adaptive tetrolet transform on image plane of preprocessed HSV color image. This transform provides the multi-resolution images with finer details by employing the tetrominoes and the proper arrangement of tetrominoes contributes the effective local geometry of image plane. The gray level co-occurrence matrix based texture feature descriptor is obtained by computing second order statistical parameters from each decomposed sub-image. At this stage, the most of the irrelevant images are discarded by retrieving the images from intermediate dataset but still some undesired images are left, those will be handled at the last stage. At this stage, fused color information is captured by applying the color autocorrelogram on both the non-uniform quantized color components of the preprocessed HSV color image. Finally, the color feature descriptor produces the desired retrieval results by discarding the irrelevant images from the texture based sub-image dataset. The proposed scheme has also low computational overhead due to the use of three descriptors at different stages separately. The retrieved results show the better accuracy as compared to the other related visual contents based image retrieval schemes.

Stationary Tetrolet Transform: an Improved Algorithm for Tetrolet Transform

In order to get an efficient image multi-scale geometrical representation, the basic principle of tetrolet transform is studied and an efficient stationary tetrolet transform algorithm based on Haar wavelet transform is proposed. Tetrolets are Haar-type wavelets whose supports are tetrominoes which are shapes made by connecting four equal-sized squares. Tetrolet transform is divide the image into 4 × 4 blocks in the horizontal and vertical direction, there is no overlap between the sub-blocks. Because there is no overlap between the images blocks, so the image decomposition coefficients in image processing is easy produce Gibbs phenomenon. Stationary tetrolet transform is a new adaptive Haar-type wavelet transforms which the block overlaps method in image decomposition process is inserted into the middle of the original coefficients in tetrolet transform. The corresponding filter bank algorithm is simple but enormously effective. Compare with standard two dimensional wavelet transform, stationary tetrolet transform is a novel tetrominoes based multi-scale geometrical transform tool, which can capture image anisotropic geometrical structures information efficiently by multi-direction selection. In this paper, the decomposition and reconstruction algorithms of the stationary tetrolet transform are described in detail, and the simulation and analysis of the decomposition of the image using the stationary tetrolet transform is carried out. Experimental results show that compared with traditional algorithm, the proposed algorithm can get better sparse representation and eliminate the blocking artifacts in image processing resulted from tetrolet transform algorithm.

Complexity reduction method for ultrasound imaging enhancement in tetrolet transform domain

Medical ultrasound images are often very obscure due to the capture devices used. As medical ultrasound images can provide information regarding potential pathological changes and abnormalities, images must be preserved and enhanced to help doctors to monitor the course of diseases. This work proposes using the energy concentration characteristics of tetrolet transform (TT) to preserve texture information and integrate the technique of adaptive histogram equalization (AHE) in order to automatically adjust the ultrasound image contrast, thereby resolving the aforementioned problems. However, the implementation of the TT in software structure brings issues such as computational complexity and large frame buffer use. In view of these problems, this work proposes a low-complexity matching pattern (LMP) method to reduce the space complexity. The LMP is combined with a look-up table approach that solves the issue of the excessive calculations generated in AHE during linear interpolation. The experimental results indicate that the modified algorithm is applicable in ultrasound imaging systems as it is able to improve calculation times.

Decision support system for glaucoma diagnosis using optical coherence tomography images

Glaucoma is an irreversible disease that damages eye's optic nerve. To avoid the vision loss, early diagnosis of glaucoma is required. In this work, an efficient Decision Support System (DSS) for glaucoma diagnosis based on Tetrolet Transform (TT) is proposed. Optical Coherence Tomography (OCT) images are used for the diagnosis. The classification of glaucomatous images is consists of three different stages; preprocessing, feature extraction and classification stage. In the preprocessing stage, Region of Inertest (ROI) is extracted from the OCT image that contains only the retinal area. In feature extraction stage, the ROI image is decomposed by TT at predefined resolution levels and statistical features are extracted from both TT decomposed image and Gray Level Difference Method (GLDM) of TT decomposed image for classification. Finally, a decision is made using Support Vector Machine (SVM) classifier. Results show that the DSS for glaucoma diagnosis provides an accuracy of 99.5% with 100% sensitivity and 99% specificity at 3rd level TT features.

Performance Comparison of Tetrolet Transform and Wavelet-Based Transforms for Medical Image Denoising

Noise reduces the quality of medical images and raise the difficulties of diagnosis. Although the wavelet transform has already been used in medical noise removal applications extensively, there are many other multi-resolution analysis methods proposed in recent years for denoising. The main goal of this study is comparing the image denoising abilities of some of these methods with wavelet transform. In this paper, image denoising is implemented by a three-stage methodology. Effectiveness of the multiresolution analysis methodologies has been investigated for standard test images beside magnetic resonans, mammography and fundus images. Performances of the transforms are compared by using peak signal to noise ratio, mean square error, mean structural similarity index and feature similarity index. The best results are obtained by tetrolet transform for random and rician noise with the benchmark images. Medical image denoising performance of Tetrolet transform is compared to other multiresolution analysis methods for the first time in the literature with this study. It surpassed ridgelet and haar wavelet transforms while the noise ratio was low. On the other hand, it is seen that curvelet transforms are effectively produce the best results for all rates of noise on medical images.

Detection of Brain Tumour in MRI Scan Images using Tetrolet Transform and SVM Classifier

Objectives: To design a system with high accuracy, sensitivity and specificity to detect brain tumour. Methods/ Statistical Analysis: Over the years, extensive research works have been done for brain tumour diagnosis using various imaging modalities. Magnetic Resonance Imaging (MRI) method has been proven to be effective in imaging the human brain for detection and classification of the tumours for clinical diagnosis. In this paper, Tetrolet Transform (TT) and Support Vector Machine (SVM) Classifier have been utilized for recognition of tumour from brain MRI images. Findings: The brain MRI image is transformed into a set of features that best describes the information content of the image. So as to use the Tetrolet transform for recognition of tumour, MRI brain tumour image is decomposed by utilizing the Tetrolet transform at a predefined level. Statistical features such as mean, standard deviation, and variance of each tetromino sub bands are computed and extracted. The t-test class separability model is applied to the calculated features keeping in mind the objective to choose the best features from the database images. The receiver operating characteristics (ROC) curves and confusion matrix are make use of, to decide the accuracy level of the proposed method. The experimental results show an accuracy level of 98.8% with fifth level of decomposition and SVM based classification while testing with images from Repository of Molecular Brain Neoplasia Data (REMBRANDT) database. The performance of the proposed system have been evaluated by using the parameters namely- accuracy, sensitivity and specificity. Application/Improvements: This technique is applied to the training images and the selected features are put away in a feature vector. 10-fold cross validation method is utilized to test viability of the proposed technique.

基于Tetrolet变换的彩色水下图像清晰化算法

基于Tetrolet变换的彩色水下图像清晰化算法

A New Transform for Medical Image Denoising: Fused Tetrolet Transform

A New Transform for Medical Image Denoising: Fused Tetrolet Transform

Surface Defect Recognition of Hot-rolled Steel Plates Based on Tetrolet Transform

Surface Defect Recognition of Hot-rolled Steel Plates Based on Tetrolet Transform

Tetrolet transform: A new adaptive Haar wavelet algorithm for sparse image representation

In order to get an efficient image representation we introduce a new adaptive Haar wavelet transform, called Tetrolet Transform. Tetrolets are Haar-type wavelets whose supports are tetrominoes which are shapes made by connecting four equal-sized squares. The corresponding fast filter bank algorithm is simple but very effective. In every level of the filter bank algorithm we divide the low-pass image into 4 × 4 blocks. Then in each block we determine a local tetrolet basis which is adapted to the image geometry in this block. An analysis of the adaptivity costs leads to modified versions of our method. Numerical results show the strong efficiency of the tetrolet transform for image approximation.