Efficient Medical Image Research Articles

Bridging 2D and 3D segmentation networks for computation-efficient volumetric medical image segmentation: An empirical study of 2.5D solutions.

Recently, deep convolutional neural networks have achieved great success for medical image segmentation. However, unlike segmentation of natural images, most medical images such as MRI and CT are volumetric data. In order to make full use of volumetric information, 3D CNNs are widely used. However, 3D CNNs suffer from higher inference time and computation cost, which hinders their further clinical applications. Additionally, with the increased number of parameters, the risk of overfitting is higher, especially for medical images where data and annotations are expensive to acquire. To issue this problem, many 2.5D segmentation methods have been proposed to make use of volumetric spatial information with less computation cost. Despite these works lead to improvements on a variety of segmentation tasks, to the best of our knowledge, there has not previously been a large-scale empirical comparison of these methods. In this paper, we aim to present a review of the latest developments of 2.5D methods for volumetric medical image segmentation. Additionally, to compare the performance and effectiveness of these methods, we provide an empirical study of these methods on three representative segmentation tasks involving different modalities and targets. Our experimental results highlight that 3D CNNs may not always be the best choice. Despite all these 2.5D methods can bring performance gains to 2D baseline, not all the methods hold the benefits on different datasets. We hope the results and conclusions of our study will prove useful for the community on exploring and developing efficient volumetric medical image segmentation methods.

An Extra Set of Intelligent Eyes: Application of Artificial Intelligence in Imaging of Abdominopelvic Pathologies in Emergency Radiology.

Imaging in the emergent setting carries high stakes. With increased demand for dedicated on-site service, emergency radiologists face increasingly large image volumes that require rapid turnaround times. However, novel artificial intelligence (AI) algorithms may assist trauma and emergency radiologists with efficient and accurate medical image analysis, providing an opportunity to augment human decision making, including outcome prediction and treatment planning. While traditional radiology practice involves visual assessment of medical images for detection and characterization of pathologies, AI algorithms can automatically identify subtle disease states and provide quantitative characterization of disease severity based on morphologic image details, such as geometry and fluid flow. Taken together, the benefits provided by implementing AI in radiology have the potential to improve workflow efficiency, engender faster turnaround results for complex cases, and reduce heavy workloads. Although analysis of AI applications within abdominopelvic imaging has primarily focused on oncologic detection, localization, and treatment response, several promising algorithms have been developed for use in the emergency setting. This article aims to establish a general understanding of the AI algorithms used in emergent image-based tasks and to discuss the challenges associated with the implementation of AI into the clinical workflow.

Multimodal Classification on PET/CT Image Fusion for Lung Cancer: A Comprehensive Survey

Medical image fusion has become essential for accurate diagnosis. For example, a lung cancer diagnosis is currently conducted with the help of multimodality image fusion to find anatomical and functional information about the tumor and metabolic measurements to identify the lung cancer stage and metastatic information of the disease. Generally, the success of multimodality imaging for lung cancer diagnosis is due to the combination of PET and CT imaging advantages while minimizing their respective limitations. However, medical image fusion involves the registration of two different modalities, which is time-consuming and technically challenging, and it is a cause of concern in a clinical setting. Therefore, the paper's main objective is to identify the most efficient medical image fusion techniques and the recent advances by conducting a collective survey. In addition, the study delves into the impact of deep learning techniques for image fusion and their effectiveness in automating the image fusion procedure with better image quality while preserving essential clinical information.

An Efficient Medical Image Processing Approach Based on a Cognitive Marine Predators Algorithm

Image processing aims to enhance the image's quality such that it is simple for both people and robots to understand. Medical image processing and Biomedical signal processing have many conceptual similarities. Medical image processing involves evaluation, enhancement, and presentation. The focus of medical imaging is on obtaining photographs for both therapeutic and diagnostic reasons. In the existing Marine Predator Algorithm, different disadvantages are experienced when various automated optimization algorithms are used to the problem of ECG categorization. The proposed method follows the flow outlined here: data collection, image preprocessing using histogram equalization, segmentation using the Otsu threshold algorithm, feature extraction using the contour method, feature selection using the Neighborhood Component Analysis (NCA) algorithm, and Cognitive Marine Predator Algorithm (CMPA) as the proposed method. By using the Cognitive Marine Predators Algorithm (CMPA), base layers are fused to use the greatest feasible parameters, producing enhanced high-quality output images. Finally, the image processing performance is analyzed. The proposed approaches overcome the drawbacks of existing algorithms and increase the quality of medical images efficiently.

An efficient content-based medical image retrieval based on a new Canny steerable texture filter and Brownian motion weighted deep learning neural network

An efficient content-based medical image retrieval based on a new Canny steerable texture filter and Brownian motion weighted deep learning neural network

A novel two phases Petri net image segmentation based on rough sets and roughness entropy

Petri net expresses concurrent events, which is very helpful for the segmentation and connection of the subsets with the image contour subsets. Combined with rough sets or roughness entropy theory for context related connections, the unexpected constraints of Petri net aiming to control precising image segmentation can be very challenging, especially in cases of sensitivity to vague and indistinct features and other inherent complicating aspects of image medical inhomogeneity. Our main contribution in this paper is to propose a two phases Petri net approach achieving forward or backward correction on multiple boundary choices, based on a rough set and roughness entropy, with the aim of more accurate and efficient medical image segmentation. The first phase in our method is to obtain an approximate contour using rough sets and roughness entropy. The second phase is to achieve a more exact contour match via unexpected constraints of a Petri net – a technique offering advantage for parallel and multiple choices applications. Experiments demonstrate that our proposed method can handle vague, uncertain and intensity inhomogeneous segmentation problems, especially where there are several subset contours needed to achieve good image segmentation.

Modelling of Efficient Medical Image Compression System Based on Joint Operation of DWT and Neural Network

Modelling of Efficient Medical Image Compression System Based on Joint Operation of DWT and Neural Network

Efficient Medical Image Encryption Framework against Occlusion Attack

Image encryption has attracted a lot of interest as an important security application for protecting confidential image data against unauthorized access. An adversary with the power to manipulate cipher image data can crop part of the image out to prevent decryption or render the decrypted image useless. This is known as the occlusion attack. In this paper, we address a vulnerability to the occlusion attack identified in the medical image encryption framework recently proposed in []. We propose adding a pixel scrambling phase to the framework and show through simulation that the extended framework effectively mitigates the occlusion attack while maintaining the other attractive security features. The scrambling is performed using a separate chaotic map which is securely initialized using a secret key and a random nonce to deter chosen-plaintext attacks. Moreover, we show through simulation that the choice of chaotic map used for scrambling is irrelevant to the effectiveness of the scrambling algorithm against the occlusion attack.

Efficient Medical Image Segmentation Based on Knowledge Distillation.

Recent advances have been made in applying convolutional neural networks to achieve more precise prediction results for medical image segmentation problems. However, the success of existing methods has highly relied on huge computational complexity and massive storage, which is impractical in the real-world scenario. To deal with this problem, we propose an efficient architecture by distilling knowledge from well-trained medical image segmentation networks to train another lightweight network. This architecture empowers the lightweight network to get a significant improvement on segmentation capability while retaining its runtime efficiency. We further devise a novel distillation module tailored for medical image segmentation to transfer semantic region information from teacher to student network. It forces the student network to mimic the extent of difference of representations calculated from different tissue regions. This module avoids the ambiguous boundary problem encountered when dealing with medical imaging but instead encodes the internal information of each semantic region for transferring. Benefited from our module, the lightweight network could receive an improvement of up to 32.6% in our experiment while maintaining its portability in the inference phase. The entire structure has been verified on two widely accepted public CT datasets LiTS17 and KiTS19. We demonstrate that a lightweight network distilled by our method has non-negligible value in the scenario which requires relatively high operating speed and low storage usage.

An efficient medical image super resolution based on piecewise linear regression strategy using domain transform filtering

SummaryHigh quality medical images are expected for the detailed analysis in medical diagnostic system. However, spatial resolution of medical images mostly suffers from the factors such as medical equipment and time constraints. Despite these limitations, a well‐designed super‐resolution (SR) algorithm will help to improve the resolution of the medical images for medical diagnosis. When comparing to upgrade medical equipment, the adaptation of SR algorithms as a post‐processing method after medical imaging gives the benefits of lower cost and superior performance. This article proposes a piecewise linear regression system for learning image specific low‐to‐high resolution mapping via domain transform filtering and weighted least squares (WLS) optimization framework. We initially employed a WLS optimization framework for gradually coarsen the original input images and extract the multi‐scale information by constructing edge‐preserving multi‐scale decompositions. Then, with the aim of adequately preserving the edges of the medical images, we utilize a recursive filtering in transform domain. The Hadamard patterns generated from low‐resolution training patches are then used to classify the LR–HR training patch pairs into different classes. In the end, the piecewise linear regression system is utilized to learn the mapping relationship between LR space to HR space for each class, which is subsequently utilized to obtain desired HR image. In the context of quantitative metrices and qualitative analysis, our proposed method generates high‐quality medical images as compares to other existing state‐of‐the‐art SR methods.

Efficient Medical Image Adaption via Axis-Aligned Mesh Deformation

Content-aware medical image adaptation can make medical images be well presented on different display devices. The existing adaption algorithms mainly consider the visual effect of salient regions, such as specific organ areas of the patient body, but either ignore the quality of unimportant areas or execute more slowly. In order to enhance the effect of adaption and accelerate the speed of adaptation, we propose an efficient medical image adaptation method via axis-aligned mesh deformation. With this method, importance map is firstly produced by combing the weighted edge map and saliency map. Then, integer programming is used to initialize and deform the axis-aligned mesh based on importance map. Finally, image adaptation is operated rapidly by bi-linear interpolation. With the proposed method, the real-time image adaptation can be realized, and not only the visual effect of the significant areas but also the contour integrity and continuity of the non significant areas can be maintained. Experiments on open data-sets show that the proposed method has high efficiency, better effect and strong stability, and is suitable for real-time image adaptation.

Efficient Medical Image Fusion Using 2-Dimensional Double Density Wavelet Transform to Improve Quality Metrics

In the recent past, a medical 2D image fusion approach has played a vital role in the biomedical industry for diagnosing disease using various modalities of medical images. Biomedical 2D image fusion synthesis is a combining technique that merges useful information from two or more modalities of data captured through high definition scanning devices. The fused image is highly useful when doctors make emergency surgical decisions for their clinical patients. In this research, an innovative medical 2D image fusion approach called 2-Dimensional Double Density Wavelet Transform (2D-DDWT), based on maximization of scaling rule, is being proposed for the combination of CT and MRI images of human brain. The source image is applied to 2D-DDWT pursued by the fusion of sub images of coefficients based on high and low frequencies that are decomposed by transformation technique. The low frequency information is decomposed using the congruency rule of phase while the high frequency information is decomposed using 2D Gabor filter. The 2D-DDWT frequency coefficients are fused by using Principle Component Analysis (PCA) for the approximation parameters, and therefore the rule of selection maximum is being applied for the coefficients to improve features of image fusion such as segmentation. Maximization scaling rule is proposed for pixel level fusion to improve quality metrics. The quantitative and qualitative analysis proves the efficiency of the proposed methodology and demonstrates the improvement of the proposed methodology over existing methods.

Robustly correlated key‐medical image for DNA‐chaos based encryption

Medical images include confidential and sensitive information about patients. Hence, ensuring the security of these images is a crucial requirement. This paper proposes an efficient and secure medical image encryption-decryption scheme based on deoxyribonucleic acid (DNA), one-dimensional chaotic maps (tent and logistic maps), and hash functions (SHA-256 and MD5). The first part of the proposed scheme is the key generation based on the hash functions of the image and its metadata. The key then is highly related and intensely sensitive to the original image. The second part is the rotation and permutation of the first two MSB bit-plans of the medical image to reduce its black background that produces redundant DNA encoded sequences. The third part is the DNA encoding-decoding using dynamically chosen DNA rules for every 2-bit pixel value through the logistic map. Meanwhile, the confusion-diffusion is performed using the tent map and XOR operation. Simulation results and security analysis prove the good encryption effects of the proposed scheme compared to the state-of-art methods with a correlation of 6.66617e-7 and a very large key space of 2624. Furthermore, the proposed system has a strong ability to resist various common attacks such as chosen/known-plaintext attacks and cropping/noise attacks.

Medical image super-resolution via deep residual neural network in the shearlet domain

This paper proposes aconvolutional neural network (CNN)-based efficient medical image super-resolution (SR) method in the shearlet domain. Because of differences between imaging mechanisms optimized for natural images and medical images, the design begins with building a medical image dataset for medical image SR and extracting effective areas to remarkably enhance the training effects of the CNN-based method. Then, a new medical image SR network structure—deep medical super-resolution network (DMSRN)—has been designed in which local residual learning is implemented through a recursive network and combined with global residual learning to heighten the depth of the network on the ground with no parameter increase. This effectively fixes the long-term dependency problem, which causes the prior state layers to barely have any effect on the following state layers. Last, the design addresses the problem of too-smooth reconstruction effects in the CNN-based method in the image space domain; shearlet transform is introduced to DMSRN to restore global topology through low-frequency sub-bands and restore local edge detail information through high-frequency sub-bands. Experimental results show that the proposed method is better than other state-of-the-art methods for medical image SR, which significantly promotes the restoration ability of texture structure and edge details.

Learning With Context Feedback Loop for Robust Medical Image Segmentation.

Deep learning has successfully been leveraged for medical image segmentation. It employs convolutional neural networks (CNN) to learn distinctive image features from a defined pixel-wise objective function. However, this approach can lead to less output pixel interdependence producing incomplete and unrealistic segmentation results. In this paper, we present a fully automatic deep learning method for robust medical image segmentation by formulating the segmentation problem as a recurrent framework using two systems. The first one is a forward system of an encoder-decoder CNN that predicts the segmentation result from the input image. The predicted probabilistic output of the forward system is then encoded by a fully convolutional network (FCN)-based context feedback system. The encoded feature space of the FCN is then integrated back into the forward system's feed-forward learning process. Using the FCN-based context feedback loop allows the forward system to learn and extract more high-level image features and fix previous mistakes, thereby improving prediction accuracy over time. Experimental results, performed on four different clinical datasets, demonstrate our method's potential application for single and multi-structure medical image segmentation by outperforming the state of the art methods. With the feedback loop, deep learning methods can now produce results that are both anatomically plausible and robust to low contrast images. Therefore, formulating image segmentation as a recurrent framework of two interconnected networks via context feedback loop can be a potential method for robust and efficient medical image analysis.

An efficient medical image encryption using hybrid DNA computing and chaos in transform domain.

In this growing era, a massive amount of digital electronic health records (EHRs) are transferred through the open network. EHRs are at risk of a myriad of security threats, to overcome such threats, encryption is a reliable technique to secure data. This paper addresses an encryption algorithm based on integer wavelet transform (IWT) blended with deoxyribo nucleic acid (DNA) and chaos to secure the digital medical images. The proposed work comprises of two phases, i.e. a two-stage shuffling phase and diffusion phase. The first stage of shuffling starts with initial block confusion followed by row and column shuffling of pixels as the second stage. The pixels of the shuffled image are circularly shifted bitwise at the first stage of diffusion to enhance the security of the system against differential attack. The second stage of diffusion operation is based on DNA coding and DNA XOR operations. The experimental analyses have been carried out with 100 DICOM test images of 16-bit depth to evaluate the strength of the algorithm against statistical and differential attacks. By the results, the maximum entropy has been obtained an average of 15.79, NPCR of 99.99, UACI of 33.31, and larger keyspace of 10140, which infer that our technique overwhelms various other state-of-the-art techniques.

UMS-Rep: Unified modality-specific representation for efficient medical image analysis.

Medical image analysis typically includes several tasks such as enhancement, segmentation, and classification. Traditionally, these tasks are implemented using separate deep learning models for separate tasks, which is not efficient because it involves unnecessary training repetitions, demands greater computational resources, and requires a relatively large amount of labeled data. In this paper, we propose a multi-task training approach for medical image analysis, where individual tasks are fine-tuned simultaneously through relevant knowledge transfer using a unified modality-specific feature representation (UMS-Rep). We explore different fine-tuning strategies to demonstrate the impact of the strategy on the performance of target medical image tasks. We experiment with different visual tasks (e.g., image denoising, segmentation, and classification) to highlight the advantages offered with our approach for two imaging modalities, chest X-ray and Doppler echocardiography. Our results demonstrate that the proposed approach reduces the overall demand for computational resources and improves target task generalization and performance. Specifically, the proposed approach improves accuracy (up to ∼ 9% ↑) and decreases computational time (up to ∼ 86% ↓) as compared to the baseline approach. Further, our results prove that the performance of target tasks in medical images is highly influenced by the utilized fine-tuning strategy.

An efficient medical image classification using deep learning techniques

Objective: Image Processing will be preparing of pictures utilizing numerical activity by utilizing some other sign handling for which the info is a picture. The yield of picture handling might be either a picture or some other structure identified with picture. Method: In therapeutic application Image Fusion is the way toward consolidating pertinent data from at least two pictures into a sign picture. The subsequent melded picture will be more instructive than any of the yield pictures. The division and order are utilized to discover the tumor part in the melded picture. Different existing multi scale change, for example, DWT, ST, DTWT, DCT produces curios in the intertwined picture. Another restorative picture combination plot dependent on a Stationary Wavelet Transform is proposed. Result: The First procedure is to apply the preprocessing technique to expel the clamor from the picture. Stationary Wavelet change is utilized to deteriorate the picture. Morphological work is utilized to extricate the component of the picture. The proposed picture gives the top notch picture when contrasted with past technique. At last, division and grouping strategies are utilized to locate the influenced zone in the melded picture. Conclusion: For Segmentation GMM (Gaussian Mixture Model) is proposed. For arrangement CNN (Convolutional Neural Networks) is proposed. In this proposed technique CT and MRI pictures are utilized as info picture to deliver progressively useful picture.

Retraction: An Efficient Deep Learning Approach for Brain Tumor Segmentation Using CNN (IOP Conf. Ser.: Mater. Sci. Eng. 981 022012)

This article has been retracted by IOP Publishing following an allegation that this article may contain tortured phrases [1]. IOP Publishing has investigated and agrees the article contains a number of nonsensical phrases that feature throughout the paper, masking overlap with previously published work [2-6], to the extent that the article makes very little sense. This casts serious doubt over the legitimacy of the article and/or expertise of the authors in this topic. IOP Publishing wishes to credit PubPeer Commenters [7] and the Problematic Paper Screener for bringing the issue to our attention. The authors agree to this retraction. [1] Cabanac G, Labbe C, Magazinov A, 2021, arXiv:2107.06751v1 [2] Karthik G.V.S, Fathima S.Y, Rahman M.Z.U, Ahamed S.R and Lay-Ekuakille A, 2013, Efficient Signal Conditioning Techniques for Brain Activity in Remote Health Monitoring Network, IEEE Sensors Journal. 13. 9 [3] Nayak S R, Mishra J and Palai G, 2018, An extended DBC approach using maximum Euclidian distance for fractal dimension of color images, Optik 166. 110-115 [4] Aparna P and Kishore P.V.V, 2018, An Efficient Medical Image Watermarking Technique in E-healthcare Application Using Hybridization of Compression and Cryptography Algorithm, Journal of Intelligent Systems. 27. 1 [5] Nithya S, Ajantha S , Mahapriya G, Mispa J, 2017, Measurement of Multimodal Brain Tumor Segmentation using Neural Network, IJERT, ICONNECT-2017 conference proceedings [6] Akshaya B, Nandhini J, Pauline Shilpa Ashwanthy J, Therasa M, 2016, Brain Tumor Segmentation Based on SFCM using Neural Network, IRJET 3 10 [7] https://pubpeer.com/publications/791A55B16A57C49057A542CA891EE9 Retraction published: 30 December 2022

A new efficient medical image cipher based on hybrid chaotic map and DNA code

In this paper, we propose a novel medical image encryption algorithm based on a hybrid model of deoxyribonucleic acid (DNA) masking, a Secure Hash Algorithm SHA-2 and a new hybrid chaotic map. Our study uses DNA sequences and operations and the chaotic hybrid map to strengthen the cryptosystem. The significant advantages of this approach consist in improving the information entropy which is the most important feature of randomness, resisting against various typical attacks and getting good experimental results. The theoretical analysis and experimental results show that the algorithm improves the encoding efficiency, enhances the security of the ciphertext, has a large key space and a high key sensitivity, and is able to resist against the statistical and exhaustive attacks.