A Privacy Preserving Hybrid Neural-Crypto Computing-Based Image Steganography for Medical Images
01 Jan 2021-pp 277-290
TL;DR: In this article, the authors proposed a two-level security for medical images using neural networks and a cryptographic algorithm (RSA) to achieve a lower loss of 0.002188 on medical images.
Abstract: With advancements in X-ray technology, there is an increase in the number of digital images used in the diagnosis of a patient. Whether it be a simple X-ray, MRI, CT scan or even a photo taken from a camera, the rise in the use of digital images has increased sharply. Though this has eased the entire process, it has brought the threat of cyber-attacks and breaches. The proposed method bridges this existing gap by incorporating suitable security mechanisms to preserve the privacy and confidentiality of medical diagnostic information of an individual. The approach utilizes neural networks to perform image steganography and combines it with a cryptographic algorithm (RSA) to secure medical images. The proposed method uses a two-level security providing a lower loss of 0.002188 on medical images improving upon existing image steganography techniques.
Citations
More filters
TL;DR: In this article , the authors present a detailed discussion of different types of medical images and the attacks that may affect medical image transmission and present an in-depth overview of security techniques, such as cryptography, steganography, and watermarking.
Abstract: Recently, there has been a rapid growth in the utilization of medical images in telemedicine applications. The authors in this paper presented a detailed discussion of different types of medical images and the attacks that may affect medical image transmission. This survey paper summarizes existing medical data security approaches and the different challenges associated with them. An in-depth overview of security techniques, such as cryptography, steganography, and watermarking are introduced with a full survey of recent research. The objective of the paper is to summarize and assess the different algorithms of each approach based on different parameters such as PSNR, MSE, BER, and NC.
16 citations
29 Mar 2022
TL;DR: The performance of the bottleneck layer on the ANN and DNN algorithm is verified in the proposed work with an openly available CIDDS-001 dataset dataset, which contains server traffic data on OpenStack and external severs.
Abstract: Cyber-attack is an attempt made from an individual or cybercriminals to hack a particular computer or network through internet. This leads to loss of information stored in the connected system and in certain cases it leads to denial of service. The traditional methods on addressing cyber-attacks are not efficient to the complex and high sophisticated attacks. Hence the deep learning based techniques are generated in recent years for estimating the attacks presence in a communication network. However, the deep learning networks are complex in nature as they are handled with a huge range of features during its operation. Therefore a bottleneck layer was developed to reduce the parameters count and feature formulations from a given data. The residual blocks are deeper than the traditional network architectures and it is achieved by enabling a 1x1 convolution block in the design flow. The performance of the bottleneck layer on the ANN and DNN algorithm is verified in the proposed work with an openly available CIDDS-001 dataset dataset. The CIDDS dataset is one of the recent dataset consists of server traffic data on OpenStack and external severs.
2 citations
01 Jan 2022
TL;DR: In this article , a new medical image cipher architecture based on fusional chaotic map is proposed, which is able to resisit Statistical, exhaustive, crop and noise attacks attacks.
Abstract: Medical images are one of the moset significant attribute for diagnoising the disease in medical systems. In today modernization in digital environment medical images are hacked during transmission on insecure network. By considering the patient’s privacy and security their medical images has to be transferred in secure maner. This work aims for proposing a new medical image cipher architecture by based on fusional chaotic map. At first, a fusional Cubic-Sine Map (CSM) is proposed to generate pseudorandom numbers, then confusion and diffusion of image is executed based on the chaotic series produced by CSM. Experimental results and security analysis indicate that the developed chaotic map model generate sufficient random series and also the proposed cipher model has the ability to resisit Statistical, exhaustive, crop and noise attacks attacks.
References
More filters
05 Oct 2015
TL;DR: Neber et al. as discussed by the authors proposed a network and training strategy that relies on the strong use of data augmentation to use the available annotated samples more efficiently, which can be trained end-to-end from very few images and outperforms the prior best method (a sliding-window convolutional network) on the ISBI challenge for segmentation of neuronal structures in electron microscopic stacks.
Abstract: There is large consent that successful training of deep networks requires many thousand annotated training samples. In this paper, we present a network and training strategy that relies on the strong use of data augmentation to use the available annotated samples more efficiently. The architecture consists of a contracting path to capture context and a symmetric expanding path that enables precise localization. We show that such a network can be trained end-to-end from very few images and outperforms the prior best method (a sliding-window convolutional network) on the ISBI challenge for segmentation of neuronal structures in electron microscopic stacks. Using the same network trained on transmitted light microscopy images (phase contrast and DIC) we won the ISBI cell tracking challenge 2015 in these categories by a large margin. Moreover, the network is fast. Segmentation of a 512x512 image takes less than a second on a recent GPU. The full implementation (based on Caffe) and the trained networks are available at http://lmb.informatik.uni-freiburg.de/people/ronneber/u-net .
49,590 citations
Posted Content•
TL;DR: It is shown that such a network can be trained end-to-end from very few images and outperforms the prior best method (a sliding-window convolutional network) on the ISBI challenge for segmentation of neuronal structures in electron microscopic stacks.
Abstract: There is large consent that successful training of deep networks requires many thousand annotated training samples. In this paper, we present a network and training strategy that relies on the strong use of data augmentation to use the available annotated samples more efficiently. The architecture consists of a contracting path to capture context and a symmetric expanding path that enables precise localization. We show that such a network can be trained end-to-end from very few images and outperforms the prior best method (a sliding-window convolutional network) on the ISBI challenge for segmentation of neuronal structures in electron microscopic stacks. Using the same network trained on transmitted light microscopy images (phase contrast and DIC) we won the ISBI cell tracking challenge 2015 in these categories by a large margin. Moreover, the network is fast. Segmentation of a 512x512 image takes less than a second on a recent GPU. The full implementation (based on Caffe) and the trained networks are available at this http URL .
19,534 citations
TL;DR: This work proposes a novel hybrid densely connected UNet (H-DenseUNet), which consists of a 2-D Dense UNet for efficiently extracting intra-slice features and a 3-D counterpart for hierarchically aggregating volumetric contexts under the spirit of the auto-context algorithm for liver and tumor segmentation.
Abstract: Liver cancer is one of the leading causes of cancer death. To assist doctors in hepatocellular carcinoma diagnosis and treatment planning, an accurate and automatic liver and tumor segmentation method is highly demanded in the clinical practice. Recently, fully convolutional neural networks (FCNs), including 2-D and 3-D FCNs, serve as the backbone in many volumetric image segmentation. However, 2-D convolutions cannot fully leverage the spatial information along the third dimension while 3-D convolutions suffer from high computational cost and GPU memory consumption. To address these issues, we propose a novel hybrid densely connected UNet (H-DenseUNet), which consists of a 2-D DenseUNet for efficiently extracting intra-slice features and a 3-D counterpart for hierarchically aggregating volumetric contexts under the spirit of the auto-context algorithm for liver and tumor segmentation. We formulate the learning process of the H-DenseUNet in an end-to-end manner, where the intra-slice representations and inter-slice features can be jointly optimized through a hybrid feature fusion layer. We extensively evaluated our method on the data set of the MICCAI 2017 Liver Tumor Segmentation Challenge and 3DIRCADb data set. Our method outperformed other state-of-the-arts on the segmentation results of tumors and achieved very competitive performance for liver segmentation even with a single model.
1,561 citations
1,220 citations
TL;DR: It is argued that steganography by itself does not ensure secrecy, but neither does simple encryption, and if these methods are combined, however, stronger encryption methods result.
Abstract: Steganography is the art of hiding information in ways that prevent the detection of hidden messages. It includes a vast array of secret communications methods that conceal the message's very existence. These methods include invisible inks, microdots, character arrangement, digital signatures, covert channels, and spread spectrum communications. Steganography and cryptography are cousins in the spycraft family: cryptography scrambles a message so it cannot be understood while steganography hides the message so it cannot be seen. In this article the authors discuss image files and how to hide information in them, and discuss results obtained from evaluating available steganographic software. They argue that steganography by itself does not ensure secrecy, but neither does simple encryption. If these methods are combined, however, stronger encryption methods result. If an encrypted message is intercepted, the interceptor knows the text is an encrypted message. But with steganography, the interceptor may not know that a hidden message even exists. For a brief look at how steganography evolved, there is included a sidebar titled "Steganography: Some History."
644 citations