This paper presents a new approach of face recognition based on wavelet network using 2D fast wavelet transform and multiresolution analysis. This approach is divided in two stages: the training stage and the recognition stage. The first consists to approximate every training face image by a wavelet network. The second consists in recognition of a new test image by comparing it to all the training faces, the distances between this test face and all images from the training set are calculated in order to identify the searched person. The usual training algorithms presents some disadvantages when the weights of the wavelet network are computed by applying the back-propagation algorithm or by direct solution which requires computing an inversion of matrix, this computation may be intensive when the learning data is too large. We present in this paper our solutions to overcome these limitations. We propose a novel learning algorithm based on the 2D Fast Wavelet Transform. Furthermore, we have increased the performances of our algorithm by introducing the Levenberg–Marquardt method to optimize the learning functions and using the Beta wavelet which has at both an analytical expression and wavelet filter bank. Extensive empirical experiments are performed to compare the proposed method with other approaches as PCA, LDA, EBGM and RBF neural network using the ORL and FERET benchmarks.

A novel approach for face recognition based on fast learning algorithm and wavelet network theory

Hard exudates in retinal images are one of the most prevalent earliest signs of diabetic retinopathy. The accurate identification of hard exudates is of increasing importance in the early detection of diabetic retinopathy. In this paper, we present a novel method to identify hard exudates from digital retinal images. A feature combination based on stationary wavelet transform (SWT) and gray level co-occurrence matrix (GLCM) is used to characterize hard exudates candidates. An optimized support vector machine (SVM) with Gaussian radial basis function is employed as a classifier. A sample dataset consisting of 50 hard exudates candidates is used for identifying hard exudates. With the optimal SVM parameters, the classification accuracy of 84%, sensitivity of 88% and specificity of 80% are obtained.

Support vector machine based method for identifying hard exudates in retinal images

Software architecture for multi-bed FDK-based reconstruction in X-ray CT scanners

Medical image quality requirements have been increasingly stringent with the recent developments of medical technology. To meet clinical diagnosis needs, an effective medical image enhancement method based on convolutional neural networks (CNNs) and frequency band broadening (FBB) is proposed. Curvelet transform is used to deal with medical data by obtaining the curvelet coefficient in each scale and direction, and the generalised cross-validation is implemented to select the optimal threshold for performing denoising processing. Meanwhile, the cycle spinning scheme is used to wipe off the visible ringing effects along the edges of medical images. Then, FBB and a new CNN model based on the retinex model are used to improve the processed image resolution. Eventually, pixel-level fusion is made between two enhanced medical images from CNN and FBB. In the authors’ study, 50 groups of medical magnetic resonance imaging, X-ray, and computed tomography images in total have been studied. The experimental results indicate that the final enhanced image using the proposed method outperforms other methods. The resolution and the edge details of the processed image are significantly enhanced, providing a more effective and accurate basis for medical workers to diagnose diseases.

https://ietresearch.onlinelibrary.wiley.com/doi/epdf/10.1049/iet-ipr.2018.6380

Efficient medical image enhancement based on CNN-FBB model

A protozoan parasite extraction scheme for digital microscopic images.

Multi-resolution system for artifact removal and edge enhancement in computerized tomography images

In this paper we report on the development and characterisation of an industrial computed tomography (ICT) system to image larger aerospace components and assemblies with sizes up to φ1000 mm, 10 m height and 2000 kg weight. The system consists of a 450 kV X-ray source, a linear detector array and a 6-axes computer numerical control (CNC) object manipulator. Mechanical as well as radiographic alignments were carried out to achieve 10 μm accuracy in mechanical motion of the objects under study. Rotate-only and translate-rotate scanning geometries were implemented for different sizes of object. Bi-directional scanning was implemented in translate-rotate geometry to reduce the scanning time by half. Corrections for centre of rotation error, ring artifacts, streak artifacts, unequal ray spacing error and cupping (beam-hardening) artifacts are implemented. A filtered back projection (FBP) algorithm is implemented for image reconstruction. The modulation transfer function (MTF) of the system is estimated using a 160 mm-diameter Perspex cylinder. The system resolution is found to be 350 μm. Details on the scanning of a few components are presented. 3D isovalue surface generation is implemented as an additional feature to aid better defect evaluation.

Industrial computed tomography system for aerospace applications: development and characterisation

Industrial Computed Tomography (ICT) system has been developed indigenously for Non Destructive Evaluation (NDE) applications. ICT system consists of 450 keV X‐ray source, 256‐channel detector array, 6‐axes mechanical manipulator, data acquisition system and reconstruction software. This system handles objects up to 1000mm (shell) diameter, 10 m height and 2000 kg weight with a spatial resolution of 1‐line pair (lp)/mm. Metallic, nonmetallic, composite hardware including assemblies of varying densities and sizes were scanned and analyzed for identification of defects such as delaminations, debonds, cracks, voids, foreign inclusions and for internal details. Various image‐processing techniques were employed for better visualization and interpretation of defects. The defects were analyzed for their location, area, depth and dimensions. The salient features of ICT system in handling a wide variety of hardware have been highlighted.

Industrial Computed Tomography (ICT) System—A Versatile Tool

Beam-hardening is an artifact, which produces false integrals if polychromatic x-ray sources are used. It is due to the photon energy dependence of the attenuation coefficient. The present work proposes an algorithm for beam-hardening correction incorporating the inherent error formula developed at IIT Kanpur. The effect of beam hardening and its removal along with inherent error is shown on both simulated and experimental data set. It is compared from the point-of-view of nearness of the corrected polychromatic projection data to the desired monochromatic projection data. The results indicate that the algorithm, proposed originally for medical applications, is giving encouraging results for non-medical objects though the physical situations are vastly different.

/pdf/poly-energetic-reconstructions-in-x-ray-ct-scanners-5caag21ycc.pdf

Poly-energetic Reconstructions in X-ray CT Scanners

X‐ray Computed Tomography (CT) is a relatively new technique developed in the late 1970’s, which enables the nondestructive visualization of the internal structure of objects. Beam hardening caused by the polychromatic spectrum is an important problem in X‐ray computed tomography (X‐CT). It leads to various artifacts in reconstruction images and reduces image quality. In the present work we are considering the Artifact Reduction in Total Hip Prosthesis CT Scan which is a problem of medical imaging. We are trying to reduce the cupping artifact induced by beam hardening as well as metal artifact as they exist in the CT scan of a human hip after the femur is replaced by a metal implant. The correction method for beam hardening used here is based on a previous work. Simulation study for the present problem includes a phantom consisting of mild steel, aluminium and perspex mimicking the photon attenuation properties of a hum hip cross section with metal implant.

Sijo N. Lukose

Papers

Multi-resolution system for artifact removal and edge enhancement in computerized tomography images

Industrial computed tomography system for aerospace applications: development and characterisation

Industrial Computed Tomography (ICT) System—A Versatile Tool

Poly-energetic Reconstructions in X-ray CT Scanners

Artifact Reduction in X‐Ray CT Images of Al‐Steel‐Perspex Specimens Mimicking a Hip Prosthesis