Fast Medical Image Research Articles

A global-frequency-domain network for medical image segmentation

The UNet series networks have been a leader in the field of medical image segmentation since their introduction. However, encoder and decoder structures of the traditional UNet series network are complex, with a large number of parameters and floating-point operations. This requires a large amount of data as support for model training, but most medical datasets only contain limited numbers of samples. To address this issue, we propose a global frequency domain UNet (GFUNet), a novel architecture for fast medical image segmentation. Inspired by recent modified Multi-Layer Perceptron(MLP)-like models, we combine Fourier Transform with UNet structure to achieve more efficient and effective encoding and decoding processes. Meanwhile, A dual-domain encoding module is designed to improve the performance of the encoder and decoder by fully used frequency domain feature. Furthermore, due to the excellent property of the Fourier Transform and its optimization, our network greatly reduces the number of parameters compared to other UNets. We evaluate GFUNet on several medical segmentation tasks, achieving improved segmentation performance compared to state-of-the-art network architectures for medical image segmentation. Compared to the original UNet, the results show that we reduce the number of parameters by 46 times, reduce computational complexity by 114 times, and improved the considerable dice score.

Pulmonary Nodule Detection and Classification Using All-Optical Deep Diffractive Neural Network.

A deep diffractive neural network (D2NN) is a fast optical computing structure that has been widely used in image classification, logical operations, and other fields. Computed tomography (CT) imaging is a reliable method for detecting and analyzing pulmonary nodules. In this paper, we propose using an all-optical D2NN for pulmonary nodule detection and classification based on CT imaging for lung cancer. The network was trained based on the LIDC-IDRI dataset, and the performance was evaluated on a test set. For pulmonary nodule detection, the existence of nodules scanned from CT images were estimated with two-class classification based on the network, achieving a recall rate of 91.08% from the test set. For pulmonary nodule classification, benign and malignant nodules were also classified with two-class classification with an accuracy of 76.77% and an area under the curve (AUC) value of 0.8292. Our numerical simulations show the possibility of using optical neural networks for fast medical image processing and aided diagnosis.

Low-Memory CNNs Enabling Real-Time Ultrasound Segmentation Towards Mobile Deployment.

Convolutional Neural Networks (CNNs), which are currently state-of-the-art for most image analysis tasks, are ill suited to leveraging the key benefits of ultrasound imaging - specifically, ultrasound's portability and real-time capabilities. CNNs have large memory footprints, which obstructs their implementation on mobile devices, and require numerous floating point operations, which results in slow CPU inference times. In this paper, we propose three approaches to training efficient CNNs that can operate in real-time on a CPU (catering to the clinical setting), with a low memory footprint, for minimal compromise in accuracy. We first demonstrate the power of 'thin' CNNs, with very few feature channels, for fast medical image segmentation. We then leverage separable convolutions to further speed up inference, reduce parameter count and facilitate mobile deployment. Lastly, we propose a novel knowledge distillation technique to boost the accuracy of light-weight models, while maintaining inference speed-up. For a negligible sacrifice in test set Dice performance on the challenging ultrasound analysis task of nerve segmentation, our final proposed model processes images at 30fps on a CPU, which is 9× faster than the standard U-Net, while requiring 420× less space in memory.

Fast Medical Image Security Using Color Channel Encryption

Abstract Evolution of digital Health-care Information System established Medical Image Security as the new contemporary research area. Most of the researchers used either image watermarking or image encryption to address medical image security. However, very few proposals focused on both issues. This paper has implemented a Fast Medial Image Security algorithm for color images that uses both watermarking and encryption of each color channel. The proposed method starts with embedding of a smoothened key image (K) and patient information over the original image (I) to generate a watermarked image (W). Then, each color channel of the watermarked image (W) is encrypted separately to produce an encrypted image (E) using the same smoothened key image (K). This image can be transmitted over the public network and the original image (I) can be achieved using decryption algorithm followed by de-watermarking using the same key image (K) at the receiver. Qualitative and quantitative results of the proposed method show good performance when compared with the existing method with high Mean, PSNR and Entropy.

A novel fast medical image segmentation scheme for anatomical scans

Medical image is the visual representation of anatomy or physiology of internal structures of the body and it is useful for clinical analysis and medical intervention. Modern medical imaging devices provide an excellent view of anatomy and physiology of internal structures of the body non-invasively. However, the usage of computers to measure, examine and determine the state of internal structures of the body with accuracy and efficiency is limited. Automated medical image segmentation techniques have wide range of utility in diagnosis, treatment planning and computer integrated surgery. These automated medical image segmentation techniques could also be used as an assisting tool to radiologists by saving their time in selecting, measuring and classifying various findings. However, automated medical image segmentation is challenging because the quality of the image is low due to the presence of noise, artefacts, partial volume effects etc., low contrast between different structures in an image and intensity variations within a region itself. This research paper focuses on fastening a region based deformable model called Chan-Vese model through various first order optimization techniques. Chan - Vese model can perform segmentation effectively even in low quality images but the limitation of Chan-Vese model is that convergence towards optimal solution is slow. The objective of this work is to fasten the convergence of Chan-Vese model towards optimal solution by using various first order optimization schemes. Chan-Vese model with proposed optimization techniques is tested with X-ray, CT and MRI images of different organs. Comparative study between traditional optimization technique used in Chan-Vese model and proposed optimization techniques has been carried out. From the comparative study, it is found that Chan-Vese model with proposed optimization schemes is efficient in terms of speedy delineation with less number of iterations and processing time. Therefore, this fastened Chan-Vese model is better suited algorithm for fast image segmentation needs such as tracking of region of interest in subsequent frames in a video.

A Fast Medical Image Super Resolution Method Based on Deep Learning Network

Low-resolution medical images can hamper medical diagnosis seriously, especially in the analysis of retina images and specifically for the detection of macula fovea. Therefore, improving the quality of medical images and speeding up their reconstruction is particularly important for expert diagnosis. To deal with this engineering problem, our paper presents a fast medical image super-resolution (FMISR) method whereby the three hidden layers to complete feature extraction is as same as the super resolution convolution neural network. It is important that a well-designed deep learning network processes images in the low resolution instead of the high-resolution space and enables the super-resolution reconstruction to be more efficient. Sub-pixel convolution layer addition and mini-network substitution in the hidden layers are critical for improving the image reconstruction speed. While the hidden layers are proposed for ensuring reconstruction quality, our FMISR framework performs significantly faster and produces a higher resolution images. As such, the technique underlying this framework presents a high potential in retinal macular examination as it provides a good platform for the segmentation of retinal images.

MRI Brain Tumour Segmentation Using Hybrid Clustering and Classification by Back Propagation Algorithm

Generally the segmentation refers, the partitioning of an image into smaller regions to identify or locate the region of abnormality. Even though image segmentation is the challenging task in medical applications, due to contrary image, local observations of an image, noise image, non uniform texture of the images and so on. Many techniques are available for image segmentation, but still it requires to introduce an efficient, fast medical image segmentation methods. This research article introduces an efficient image segmentation method based on K means clustering integrated with a spatial Fuzzy C means clustering algorithms. The suggested technique combines the advantages of the two methods. K means segmentation requires minimum computation time, but spatial Fuzzy C means provides high accuracy for image segmentation. The performance of the proposed method is evaluated in terms of accuracy, PSNR and processing time. It also provides good implementation results for MRI brain image segmentation with high accuracy and minimal execution time. After completing the segmentation the of abnormal part of the input MRI brain image, it is compulsory to classify the image is normal or abnormal. There are many classifiers like a self organizing map, Back propagation algorithm, support vector machine etc., The algorithm helps to classify the abnormalities like benign or malignant brain tumour in case of MRI brain image. The abnormality is detected based on the extracted features from an input image. Discrete wavelet transform helps to find the hidden information from the MRI brain image. The extracted features are trained by Back Propagation Algorithm to classify the abnormalities of MRI brain image.

Fast medical image registration using bidirectional empirical mode decomposition

This paper focuses on an acceleration of the mutual information maximization method for medical image registration. Our approach is based on fast adaptive bidirectional empirical mode decomposition (FABEMD). The registration is performed for the informative intrinsic image modes. It aims to reduce the computational complexity of the mutual entropy maximization algorithm by extracting only essential data. Optimization process consists of several steps: image structural reduction using FABEMD, sequential parameters search, image downsampling, and, finally, multilevel parametric space search. We compare our approach to standard mutual information maximization method (MMI) and analyze results for multimodal medical images. Experiments show that proposed method produces consistent results very close to MMI, while reducing the registration time by 200 time on average.

A fast medical image segmentation method of based on initial contour forecast segmentation model

A fast segmentation algorithm of single medical image and sequence images based on active contour model are proposed in this paper. We give out the initial contour forecast segmentation model of 3D medical image first, and then, numerical solution of the image segmentation model algorithm is presented, send data blocks to different processors to do numerical calculations, finally combine the segmentation results. At the same time, parallel designing methods of node allocation, task allocation and load balancing methods are researched. The experiment results show that the parallel method can significantly improve the segmentation speed of 3D medical image.

Accurate and fast deformable medical image registration for brain tumor resection using image-guided neurosurgery

We present a parallel adaptive physics-based non-rigid registration framework for aligning pre-operative to intra-operative brain magnetic resonance images (MRI) of patients who have undergone a tumor resection. This framework extends our earlier work on the physics-based methods by using an adaptive, multi-material, parallel finite element biomechanical model to describe the physical deformations of the brain. Our registration technology incorporates fast image-to-mesh convertors for remeshing the brain model in real-time eliminating the poor-quality elements; various linear solvers to accurately estimate the volumetric deformations; efficient block-matching techniques to compensate for the missing/unrealistic matches induced by the tumor resection. Our evaluation is based on six clinical volume MRI data-sets including (i) isotropic and anisotropic image spacings, and (ii) partial and complete tumor resections. We compare our framework with four methods: a rigid and BSpline deformable registration implemented on 3D Slicer v4.4.0, a physics-based non-rigid registration available on ITK v4.7.0, and an adaptive physics-based non-rigid registration. We show that the proposed technology provides the finest MRI alignments among all the methods. The Hausdorff distance is on average up to 3.78 and 3.12 times more accurate compared to the rigid and the other non-rigid registration methods, respectively. Additionally, it brings the end-to-end execution within the real-time constraints imposed by the neurosurgical procedure. In a Linux Dell workstation with 12 Intel Xeon 3.47 GHz CPU cores and 96 GB of RAM, it registers the anisotropic volume data in less than 93 s and the isotropic data in less than 21 s.

The Fast Medical Image Segmentation of Target Region Based on Improved FM Algorithm

This paper studies several kinds of image segmentation algorithm,and region-growing algorithm and fast level set matching algorithm FM are programmed by VC and verified,thereinto,the speed of segmentation of region-growing algorithm is fast.It is primarily affected by the identity of gray level of object region, for the inconsistent object region, excessive segmentation and missing segmentation will happen.The fast matching method can easily handle the geometric objects which topological structure is complex or changing, but the evolving curve also easily leak from the boundary, if there are holes in an object which has been segmented, it will not quite separate the interior outline of the object,so,aiming at the characteristics of medical image,an improved fast matching algorithm is presented in this paper, it can effectively enhance the image segmentation effect and prevent the loss of details of lines, and the internal change of topological structures of the objects can also be segmented better by it.

Fast 3D medical image rigid registration using the mutual information in equivalent meridian plane

Mutual information (MI) technique is an effective similarity measure for the rigid registration of multimodality medical images. However, when the MI methods work with the entire image data, the computation cost is too expensive and are prone to fall into local minima. In this paper, a new concept of equivalent meridian plane (EMP) is proposed, and is determined using principal component analysis (PCA); the rough registrations of EMP is to be realized at six freedom degrees; finally, the refine registrations can be completed using MI in EMP. This method is called as EMP-based MI registration technique, which combines the feature and intensity information. The accuracy and robustness of EMP-MI approach can be verified by applying it to the simulated and real medical image data. The experimental results indicate that the proposed algorithm reduces computational time distinctly and is more robust than transitional MI technology.

A Fast Medical 3D Image Registration Method Based on a Parametric Representation

Three dimensional medical image registration is a fundamental technique which applied various medical treatments such as image diagnosis, treatment planning, image guided surgery, etc. In radiation therapy for cancer treatment, quick alignment method around target lesion is required. To align CT images, similarity measurement like normalized cross correlation or mutual information should be calculated. The computational cost must reduce for achieving the online image registration.In this paper, we propose a novel quick rigid registration method which makes it possible to align the three dimensional images using the parametric eigenspace method. By projecting each CT slice image into the eigenspace as a set of low dimensional vectors, image similarity can be calculated very rapidly. The experiments using CT images of the same patient, it is found that the alignment accuracy is almost the same as the method using normalized cross correlation, and the computation time is less than one second.