No abstract

Ultrasound images often exhibit poor signal to noise ratio when compared with optical images, because of the presence multiplicative speckle noise. Speckle suppression is often carried out as a pre-processing step to aid in diagnosis using ultrasound images. In this study, dual tree complex wavelet transform-based Levy Shrink algorithm is proposed for denoising ultrasound images. The coefficients in each wavelet subband are modelled using a heavy tailed Levy distribution. The scale parameters of the Levy distribution are estimated using fractional moments. Within this framework, a Bayesian estimator is employed to denoise ultrasound images. The proposed Levy Shrink algorithm is verified using evaluation parameters such as peak signal to noise ratio, mean structural similarity index, correlation coefficient and equivalent number of looks. The efficiency of the proposed denoising algorithm is justified by conducting extensive experiments on real as well as simulated ultrasound images.

No abstract

Automatic detection of sustained intensity changes (edges) in SAR images for applications requiring localization and identification of objects is complicated by the nature of the speckle. The speckle is characterized by a high degree of correlation and multiplicative signal dependence. Ratio based detectors like MRoA and RGoA edge detector use predefined thresholds. The modified RGoA detector defines an automatic threshold determining method. But all these edge detectors, apart from detecting the object edges, detect a number of false edges. In this paper, a new ROEWA based algorithm that automatically discriminates the object boundaries and the false edges is proposed. Otsu's nonparametric and unsupervised principle for automatic threshold selection is introduced in this classification process. An optimal threshold is selected by maximizing the seperability of the classes in gray level. Real SAR images are used to verify our method and the results are compared with the existing methods for edge detection. Experimental results show that the proposed method is robust and efficient.

Increased growth of information technology in healthcare has led to a situation where the security of patient information is more important and is a critical issue. The aim of the proposed algorithm is to provide a framework to verify the integrity of the medical images. In this paper, the integrity of the medical images is verified by embedding hash signatures using the sequential square embedding technique. This technique is as efficient as the diamond encoding technique but with increased payload capability. The medical image is first divided into the region of interest (ROI) block and the signature block. The hash signatures are determined by dividing the ROI into nonoverlapping blocks. During the data hiding stage, the hash signatures are embedded in randomly chosen pixel pairs in the signature block using the sequential square encoding (SSE) technique. In the experimental results, the data hiding capacity of the proposed SSE technique is verified in terms of peak signal-to-noise ratio. Also, the medical image integrity is substantiated by comparing the L2 norm between computed and extracted hash signatures. Modifications such as contrast enhancement, rotation, scaling, and changing the image information result in increased L2 norm; thus, the integrity of the medical images can be verified. The parameters required for embedding, such as the embedding parameter and the seed for random sequence generation, are encrypted and communicated to the receiving end. Hence, the proposed algorithm provides a secure framework for medical image integrity verification.

No abstract

No abstract

Enhancement of low-light images is a challenging task due to the impact of low brightness, low contrast, and high noise. The inability to collect natural labeled data intensifies this problem further. Many researchers have attempted to solve this problem using learning-based approaches; however, most models ignore the impact of noise in low-lit images. In this paper, an encoder-decoder architecture, made up of separable convolution layers that solve the issues encountered in low-light image enhancement, is proposed. The architecture is trained end-to-end on a custom low-light image dataset (LID), comprising both clean and noisy images. We introduce a unique multi-context feature extraction module (MC-FEM) where the input first passes through a feature pyramid of dilated separable convolutions for hierarchicalcontext feature extraction followed by separable convolutions for feature compression. The model is optimized using a novel three-part loss function that focuses on high-level contextual features, structural similarity, and patch-wise local information. We conducted several ablation studies to determine the optimal model for low-light image enhancement under noisy and noiseless conditions. We have used performance metrics like peak-signal-to-noise ratio, structural similarity index matrix, visual information fidelity, and average brightness to demonstrate the superiority of the proposed work against the state-of-the-art algorithms. Qualitative results presented in this paper prove the strength and suitability of our model for real-time applications. INDEX TERMSEncoder-decoder Architecture, Separable Convolution, Dilated Convolution, ASPP, Perceptual loss, Low-light Image Enhancement I. INTRODUCTION L OW light image enhancement is an active area of research that enables the acquisition system to capture superior quality images even under low-light conditions. Its applications include autonomous driving, photography, military, object detection, and surveillance. Low-light image enhancement (LIE) algorithms consider several factors like color contrast, brightness, image resolution, and dynamic range. Several researchers use image processing techniques to enhance low-light images, such as histogram equalization (HE) [1] that tend to equalize the dynamic range of intensities. Also, histogram equalization methods focus only on increasing the image contrast, whereas they fail to address the actual illumination issues. As mentioned in [2], linear and non-linear low-light image enhancement methods have simple yet fast implementation but don't consider the image distribution, leading to limited enhancement ability. Spatial filters and other image processing techniques have shown 18 improvement in the low-lit image quality, but these methods 19 fail to work under noisy context. Noise, typically, is amplified 20 by the filters, as they rely on a small neighborhood. These 21 reasons justify the need for deep-learning in enhancing low-22 light images, as these methods can dynamically learn and 23 handle noisy as well as cl...