Authenticating medical images using watermarking techniques has become a very popular area of research, and some works in this area have been reported worldwide recently. Besides authentication, many data-hiding techniques have been proposed to conceal patient’s data into medical images aiming to reduce the cost needed to store data and the time needed to transmit data when required. In this paper, we present a new hybrid watermarking scheme for DICOM images. In our scheme, two well-known techniques are combined to gain the advantages of both and fulfill the requirements of authentication and data hiding. The scheme divides the images into two parts, the region of interest (ROI) and the region of non-interest (RONI). Patient’s data are embedded into ROI using a reversible technique based on difference expansion, while tamper detection and recovery data are embedded into RONI using a robust technique based on discrete wavelet transform. The experimental results show the ability of hiding patient’s data with a very good visual quality, while ROI, the most important area for diagnosis, is retrieved exactly at the receiver side. The scheme also shows some robustness against certain levels of salt and pepper and cropping noise.
Watermarking; Data hiding; Medical Image Authentication; Electronic patient record
Digital medical images are very easy to be modified for illegal purposes. For example, microcalcification in mammography is an important diagnostic clue, and it can be wiped off intentionally for insurance purposes or added intentionally into a normal mammography. In this paper, we proposed two methods to tamper detection and recovery for a medical image. A 1024 × 1024 x-ray mammogram was chosen to test the ability of tamper detection and recovery. At first, a medical image is divided into several blocks. For each block, an adaptive robust digital watermarking method combined with the modulo operation is used to hide both the authentication message and the recovery information. In the first method, each block is embedded with the authentication message and the recovery information of other blocks. Because the recovered block is too small and excessively compressed, the concept of region of interest (ROI) is introduced into the second method. If there are no tampered blocks, the original image can be obtained with only the stego image. When the ROI, such as microcalcification in mammography, is tampered with, an approximate image will be obtained from other blocks. From the experimental results, the proposed near-lossless method is proven to effectively detect a tampered medical image and recover the original ROI image. In this study, an adaptive robust digital watermarking method combined with the operation of modulo 256 was chosen to achieve information hiding and image authentication. With the proposal method, any random changes on the stego image will be detected in high probability.
Medical image; image processing; image authentication
Nowadays, medical imaging equipments produce digital form of medical images. In a modern health care environment, new systems such as PACS (picture archiving and communication systems), use the digital form of medical image too. The digital form of medical images has lots of advantages over its analog form such as ease in storage and transmission. Medical images in digital form must be stored in a secured environment to preserve patient privacy. It is also important to detect modifications on the image. These objectives are obtained by watermarking in medical image.
In this paper, we present a dual and oblivious (blind) watermarking scheme in the contourlet domain. Because of importance of ROI (region of interest) in interpretation by medical doctors rather than RONI (region of non-interest), we propose an adaptive dual watermarking scheme with different embedding strength in ROI and RONI. We embed watermark bits in singular value vectors of the embedded blocks within lowpass subband in contourlet domain.
The values of PSNR (peak signal-to-noise ratio) and SSIM (structural similarity measure) index of ROI for proposed DICOM (digital imaging and communications in medicine) images in this paper are respectively larger than 64 and 0.997. These values confirm that our algorithm has good transparency. Because of different embedding strength, BER (bit error rate) values of signature watermark are less than BER values of caption watermark. Our results show that watermarked images in contourlet domain have greater robustness against attacks than wavelet domain. In addition, the qualitative analysis of our method shows it has good invisibility.
The proposed contourlet-based watermarking algorithm in this paper uses an automatically selection for ROI and embeds the watermark in the singular values of contourlet subbands that makes the algorithm more efficient, and robust against noise attacks than other transform domains. The embedded watermark bits can be extracted without the original image, the proposed method has high PSNR and SSIM, and the watermarked image has high transparency and can still conform to the DICOM format.
Tamper localization and recovery watermarking scheme can be used to detect manipulation and recover tampered images. In this paper, a tamper localization and lossless recovery scheme that used region of interest (ROI) segmentation and multilevel authentication was proposed. The watermarked images had a high average peak signal-to-noise ratio of 48.7 dB and the results showed that tampering was successfully localized and tampered area was exactly recovered. The usage of ROI segmentation and multilevel authentication had significantly reduced the time taken by approximately 50 % for the tamper localization and recovery processing.
Lossless compression; Tamper localization; Recovery; Medical image; Watermarking
Teleradiology applications and universal availability of patient records using web-based technology are rapidly gaining importance. Consequently, digital medical image security has become an important issue when images and their pertinent patient information are transmitted across public networks, such as the Internet. Health mandates such as the Health Insurance Portability and Accountability Act require healthcare providers to adhere to security measures in order to protect sensitive patient information. This paper presents a fully reversible, dual-layer watermarking scheme with tamper detection capability for medical images. The scheme utilizes concepts of public-key cryptography and reversible data-hiding technique. The scheme was tested using medical images in DICOM format. The results show that the scheme is able to ensure image authenticity and integrity, and to locate tampered regions in the images.
Digital watermark; security; image authentication; teleradiology; public-key cryptography
This paper presents a lossless watermarking scheme in the sense that the original image can be exactly recovered from the watermarked one, with the purpose of verifying the integrity and authenticity of medical images. In addition, the scheme has the capability of not introducing any embedding-induced distortion in the region of interest (ROI) of a medical image. Difference expansion of adjacent pixel values is employed to embed several bits. A region of embedding, which is represented by a polygon, is chosen intentionally to prevent introducing embedding distortion in the ROI. Only the vertex information of a polygon is transmitted to the decoder for reconstructing the embedding region, which improves the embedding capacity considerably. The digital signature of the whole image is embedded for verifying the integrity of the image. An identifier presented in electronic patient record (EPR) is embedded for verifying the authenticity by simultaneously processing the watermarked image and the EPR. Combining with fingerprint system, patient’s fingerprint information is embedded into several image slices and then extracted for verifying the authenticity.
Watermarking; telemedicine; security; integrity; confidentiality; image authentication; PACS; ROI
Given the ease of alteration of digital data, integrity verification and tamper detection for medical images are becoming ever more important. In this paper, instead of using the conventional irreversible block-based watermarking approach to achieve tamper localization, we propose to incorporate such functionality into the region-based lossless watermarking scheme. This is achieved by partitioning an image into certain non-overlapping regions and appending the associated local authentication information directly into the watermark payload. A region of authentication, which can be flexibly specified by the user, is partitioned into small regions in a multilevel hierarchical manner. Such hierarchical structure allows the user to easily adjust the localization accuracy, and makes the tamper detection efficient. Experimental results demonstrate the effectiveness of tamper localization.
Watermarking; security; integrity; image authentication; tamper localization; telemedicine; PACS; ROI
The aim of this study was to explore the technical feasibility of T1ρ MRI for the liver, and to determine the normal range of liver T1ρ in healthy subjects at clinical 3 T.
There were 15 healthy volunteers. Three representative axial slices were selected to cut through the upper, middle and lower liver. A rotary echo spin-lock pulse was implemented in a two-dimensional fast-field echo sequence. Spin-lock frequency was 500 Hz, and the spin-lock times of 1, 10, 20, 30, 40 and 50 ms were used for T1ρ mapping. The images were acquired slice by slice during breath-holding. Regions of interest (ROIs; n=5) were manually placed on each slice of the liver parenchyma region, excluding artefacts and vessels. The mean value of these ROIs (n=15) was regarded as the liver T1ρ value for the subject. Six subjects were scanned once at fasting status; six subjects were scanned once 2 h post meal; three subjects were scanned twice at fasting status; and seven subjects were scanned twice 2 h post meal.
When two readers measured the same 10 data sets, the interreader reproducibility (ICC: intraclass correlation coefficient) was 0.955. With the 10 subjects scanned twice, the ICC for scan–rescan reproducibility was 0.764. There was no significant difference for the liver T1ρ value at the fasting status (43.08±1.41 ms) and post-meal status (42.97±2.38 ms, p=0.867). Pooling together all the 32 scans in this study, the normal liver T1ρ value ranged from 38.6 to 48.3 ms (mean 43.0 ms, median 42.6 ms).
It is feasible to obtain consistent liver T1ρ measurement for human subjects at 3 T.
Identification of regions of interest (ROIs) is a fundamental issue in brain network construction and analysis. Recent studies demonstrate that multimodal neuroimaging approaches and joint analysis strategies are crucial for accurate, reliable and individualized identification of brain ROIs. In this paper, we present a novel approach of visual analytics and its open-source software for ROI definition and brain network construction. By combining neuroscience knowledge and computational intelligence capabilities, visual analytics can generate accurate, reliable and individualized ROIs for brain networks via joint modeling of multimodal neuroimaging data and an intuitive and real-time visual analytics interface. Furthermore, it can be used as a functional ROI optimization and prediction solution when fMRI data is unavailable or inadequate. We have applied this approach to an operation span working memory fMRI/DTI dataset, a schizophrenia DTI/resting state fMRI (R-fMRI) dataset, and a mild cognitive impairment DTI/R-fMRI dataset, in order to demonstrate the effectiveness of visual analytics. Our experimental results are encouraging.
multimodal neuroimaging; joint modeling; visual analytics; visualization and interaction; brain networks
Studying structural and functional connectivities of human cerebral cortex has drawn significant interest and effort recently. A fundamental and challenging problem arises when attempting to measure the structural and/or functional connectivities of specific cortical networks: how to identify and localize the best possible regions of interests (ROIs) on the cortex? In our view, the major challenges come from uncertainties in ROI boundary definition, the remarkable structural and functional variability across individuals and high nonlinearities within and around ROIs. In this paper, we present a novel ROI prediction framework that localizes ROIs in individual brains based on their learned fiber shape models from multimodal task-based functional magnetic resonance imaging (fMRI) and diffusion tensor imaging (DTI) data. In the training stage, shape models of white matter fibers are learnt from those emanating from the functional ROIs, which are activated brain regions detected from task-based fMRI data. In the prediction stage, functional ROIs are predicted in individual brains based only on DTI data. Our experiment results show that the average ROI prediction error is around 3.94 mm, in comparison with benchmark data provided by working memory and visual task-based fMRI. Our work demonstrated that fiber bundle shape models derived from DTI data are good predictors of functional cortical ROIs.
brain network; diffusion tensor imaging; fMRI; ROI prediction; shape analysis
Previous studies using hierarchical clustering approach to analyze resting-state fMRI data were limited to a few slices or regions-of-interest (ROIs) after substantial data reduction.
To develop a framework that can perform voxel-wise hierarchical clustering of whole-brain resting-state fMRI data from a group of subjects.
Materials and Methods
Resting-state fMRI measurements were conducted for 86 adult subjects using a single-shot echo-planar imaging (EPI) technique. After pre-processing and co-registration to a standard template, pair-wise cross-correlation coefficients (CC) were calculated for all voxels inside the brain and translated into absolute Pearson's distances after imposing a threshold CC≥0.3. The group averages of the Pearson's distances were then used to perform hierarchical clustering with the developed framework, which entails gray matter masking and an iterative scheme to analyze the dendrogram.
With the hierarchical clustering framework, we identified most of the functional connectivity networks reported previously in the literature, such as the motor, sensory, visual, memory, and the default-mode functional networks (DMN). Furthermore, the DMN and visual system were split into their corresponding hierarchical sub-networks.
It is feasible to use the proposed hierarchical clustering scheme for voxel-wise analysis of whole-brain resting-state fMRI data. The hierarchical clustering result not only confirmed generally the finding in functional connectivity networks identified previously using other data processing techniques, such as ICA, but also revealed directly the hierarchical structure within the functional connectivity networks.
This paper presents an adaptive attention window (AAW)-based microscopic cell nuclei segmentation method. For semantic AAW detection, a luminance map is used to create an initial attention window, which is then reduced close to the size of the real region of interest (ROI) using a quad-tree. The purpose of the AAW is to facilitate background removal and reduce the ROI segmentation processing time. Region segmentation is performed within the AAW, followed by region clustering and removal to produce segmentation of only ROIs. Experimental results demonstrate that the proposed method can efficiently segment one or more ROIs and produce similar segmentation results to human perception. In future work, the proposed method will be used for supporting a region-based medical image retrieval system that can generate a combined feature vector of segmented ROIs based on extraction and patient data.
Microscopic image; nuclei segmentation; region of interest (ROI); adaptive attention window (AAW); quad-tree; region-based image retrieval
The experimental study presented in this paper is aimed at the development of an automatic image segmentation system for classifying region of interest (ROI) in medical images which are obtained from different medical scanners such as PET, CT, or MRI. Multiresolution analysis (MRA) using wavelet, ridgelet, and curvelet transforms has been used in the proposed segmentation system. It is particularly a challenging task to classify cancers in human organs in scanners output using shape or gray-level information; organs shape changes throw different slices in medical stack and the gray-level intensity overlap in soft tissues. Curvelet transform is a new extension of wavelet and ridgelet transforms which aims to deal with interesting phenomena occurring along curves. Curvelet transforms has been tested on medical data sets, and results are compared with those obtained from the other transforms. Tests indicate that using curvelet significantly improves the classification of abnormal tissues in the scans and reduce the surrounding noise.
The study aimed at comparing PET/MR to PET/CT for imaging the carotid arteries in patients with known increased risk of atherosclerosis. Six HIV-positive men underwent sequential PET/MR and PET/CT of the carotid arteries after injection of 400 MBq of 18F-FDG. PET/MR was performed a median of 131 min after injection. Subsequently,PET/CT was performed. Regions of interest (ROI) were drawn slice by slice to include the carotid arteries and standardized uptake values (SUV) were calculated from both datasets independently. Quantitative comparison of 18F-FDG uptake revealed a high congruence between PET data acquired using the PET/MR system compared to the PET/CT system. The mean difference for SUVmean was -0.18 (p < 0.001) and -0.14 for SUVmax (p < 0.001) indicating a small but significant bias towards lower values using the PET/MR system. The 95% limits of agreement were -0.55 to 0.20 for SUVmean and -0.93 to 0.65 for SUVmax. The image quality of the PET/MR allowed for delineation of the carotid vessel wall. The correlations between 18F-FDG uptake from ROI including both vessel wall and vessel lumen to ROI including only the wall were strong (r = 0.98 for SUVmean and r = 1.00 for SUVmax) indicating that the luminal 18F-FDG content had minimal influence on the values. The study shows for the first time that simultaneous PET/MR of the carotid arteries is feasible in patients with increased risk of atherosclerosis. Quantification of 18F-FDG uptake correlated well between PET/MR and PET/CT despite difference in method of PET attenuation correction, reconstruction algorithm, and detector technology.
Atherosclerosis; positron emission tomography; magnetic resonance imaging; PET/MR; hybrid scanners
Digital mammogram has become the most effective technique for early breast cancer detection modality. Digital mammogram takes an electronic image of the breast and stores it directly in a computer. The aim of this study is to develop an automated system for assisting the analysis of digital mammograms. Computer image processing techniques will be applied to enhance images and this is followed by segmentation of the region of interest (ROI). Subsequently, the textural features will be extracted from the ROI. The texture features will be used to classify the ROIs as either masses or non-masses. In this study normal breast images and breast image with masses used as the standard input to the proposed system are taken from Mammographic Image Analysis Society (MIAS) digital mammogram database. In MIAS database, masses are grouped into either spiculated, circumscribed or ill-defined. Additional information includes location of masses centres and radius of masses. The extraction of the textural features of ROIs is done by using gray level co-occurrence matrices (GLCM) which is constructed at four different directions for each ROI. The results show that the GLCM at 0º, 45º, 90º and 135º with a block size of 8X8 give significant texture information to identify between masses and non-masses tissues. Analysis of GLCM properties i.e. contrast, energy and homogeneity resulted in receiver operating characteristics (ROC) curve area of Az = 0.84 for Otsu’s method, 0.82 for thresholding method and Az = 0.7 for K-mean clustering. ROC curve area of 0.8-0.9 is rated as good results. The authors’ proposed method contains no complicated algorithm. The detection is based on a decision tree with five criterions to be analysed. This simplicity leads to less computational time. Thus, this approach is suitable for automated real-time breast cancer diagnosis system.
Breast cancer; Mammogram; Masses; GLCM
Intervertebral spacers for anterior spine fusion are made of different materials, such as titanium, carbon or cobalt-chrome, which can affect the post-fusion MRI scans. Implant-related susceptibility artifacts can decrease the quality of MRI scans, thwarting proper evaluation. This cadaver study aimed to demonstrate the extent that implant-related MRI artifacting affects the post-fusion evaluation of intervertebral spacers. In a cadaveric porcine spine, we evaluated the post-implantation MRI scans of three intervertebral spacers that differed in shape, material, surface qualities and implantation technique. A spacer made of human cortical bone was used as a control. The median sagittal MRI slice was divided into 12 regions of interest (ROI). No significant differences were found on 15 different MRI sequences read independently by an interobserver-validated team of specialists (P>0.05). Artifact-affected image quality was rated on a score of 0-1-2. A maximum score of 24 points (100%) was possible. Turbo spin echo sequences produced the best scores for all spacers and the control. Only the control achieved a score of 100%. The carbon, titanium and cobalt-chrome spacers scored 83.3, 62.5 and 50%, respectively. Our scoring system allowed us to create an implant-related ranking of MRI scan quality in reference to the control that was independent of artifact dimensions. The carbon spacer had the lowest percentage of susceptibility artifacts. Even with turbo spin echo sequences, the susceptibility artifacts produced by the metallic spacers showed a high degree of variability. Despite optimum sequencing, implant design and material are relevant factors in MRI artifacting.
Intervertebral spacers; Implant materials; MRI; Susceptibility artifacts
Region of Interest (ROI) extraction is a crucial step in an automatic finger vein recognition system. The aim of ROI extraction is to decide which part of the image is suitable for finger vein feature extraction. This paper proposes a finger vein ROI extraction method which is robust to finger displacement and rotation. First, we determine the middle line of the finger, which will be used to correct the image skew. Then, a sliding window is used to detect the phalangeal joints and further to ascertain the height of ROI. Last, for the corrective image with certain height, we will obtain the ROI by using the internal tangents of finger edges as the left and right boundary. The experimental results show that the proposed method can extract ROI more accurately and effectively compared with other methods, and thus improve the performance of finger vein identification system. Besides, to acquire the high quality finger vein image during the capture process, we propose eight criteria for finger vein capture from different aspects and these criteria should be helpful to some extent for finger vein capture.
finger vein image; ROI extraction; sliding window; phalangeal joint; capture criteria
Research in recent years has provided some evidence of temporal non-stationarity of functional connectivity in resting state fMRI. In this paper, we present a novel methodology that can decode connectivity dynamics into a temporal sequence of hidden network “states” for each subject, using a Hidden Markov Modeling (HMM) framework. Each state is characterized by a unique covariance matrix or whole-brain network. Our model generates these covariance matrices from a common but unknown set of sparse basis networks, which capture the range of functional activity co-variations of regions of interest (ROIs). Distinct hidden states arise due to a variation in the strengths of these basis networks. Thus, our generative model combines a HMM framework with sparse basis learning of positive definite matrices. Results on simulated fMRI data show that our method can effectively recover underlying basis networks as well as hidden states. We apply this method on a normative dataset of resting state fMRI scans. Results indicate that the functional activity of a subject at any point during the scan is composed of combinations of overlapping task-positive/negative pairs of networks as revealed by our basis. Distinct hidden temporal states are produced due to a different set of basis networks dominating the covariance pattern in each state.
resting state fMRI; functional connectivity; temporal network dynamics
To assess the influence of region of interest (ROI) size and positioning on tumour ADC measurements and interobserver variability in patients with locally advanced rectal cancer (LARC).
Forty-six LARC patients were retrospectively included. Patients underwent MRI including DWI (b0,500,1000) before and 6–8 weeks after chemoradiation (CRT). Two readers measured mean tumour ADCs (pre- and post-CRT) according to three ROI protocols: whole-volume, single-slice or small solid samples. The three protocols were compared for differences in ADC, SD and interobserver variability (measured as the intraclass correlation coefficient; ICC).
ICC for the whole-volume ROIs was excellent (0.91) pre-CRT versus good (0.66) post-CRT. ICCs were 0.53 and 0.42 for the single-slice ROIs versus 0.60 and 0.65 for the sample ROIs. Pre-CRT ADCs for the sample ROIs were significantly lower than for the whole-volume or single-slice ROIs. Post-CRT there were no significant differences between the whole-volume ROIs and the single-slice or sample ROIs, respectively. The SDs for the whole-volume and single-slice ROIs were significantly larger than for the sample ROIs.
ROI size and positioning have a considerable influence on tumour ADC values and interobserver variability. Interobserver variability is worse after CRT. ADCs obtained from the whole tumour volume provide the most reproducible results.
• ROI size and positioning influence tumour ADC measurements in rectal cancer
• ROI size and positioning influence interobserver variability of tumour ADC measurements
• ADC measurements of the whole tumour volume provide the most reproducible results
• Tumour ADC measurements are more reproducible before, rather than after, chemoradiation treatment
• Variations caused by ROI size and positioning should be taken into account when using ADC as a biomarker for tumour response
Diffusion magnetic resonance imaging; Rectal neoplasms; Observer variation; Methodology; Apparent diffusion coefficient
High-resolution computed tomography (CT) reconstructions currently require either full field of view (FOV) exposure, resulting in high dose, or region of interest (ROI) exposure, resulting in artifacts. To obtain high-resolution 3D reconstruction of an ROI with minimal artifacts, we have developed a method involving a non-uniform ROI beam filter to reduce dose outside the ROI while acquiring the ROI at a higher dose. High-resolution, high-dose full-field projections of a phantom were obtained. ROIs in the images were selected and the low-dose data outside the ROI were simulated by adding various levels of noise to the projection data corresponding to a dose of 1/16 and 1/256 of the original dose. For an ROI of 30% FOV, artifacts in the reconstructed ROI were minimal for both dose reduction levels. For an ROI of 10% FOV, artifacts remained minimal only for the 1/16th dose case. The effect of the presence of a high contrast object outside the ROI was also studied. We found that the intensity of the artifacts increases with the contrast of the object, its size, and its distance from the axis of rotation. CT using an ROI filter provides a way to reconstruct an ROI with reduced integral dose and yet with minimal artifacts and improved spatial resolution.
Region of interest CT; ROI CT; Volume of interest CT; VOI CT; Area of interest CT; Artifact reduction; Dose reduction; reduced FOV; truncated reconstruction
We describe a cardiac gated high in-plane resolution axial human cervical spinal cord diffusion tensor imaging (DTI) protocol. Multiple steps were taken to optimize both image acquisition and image processing. The former includes slice-by-slice cardiac triggering and individually tiltable slices. The latter includes (i) iterative 2D retrospective motion correction, (ii) image intensity outlier detection to minimize the influence of physiological noise, (iii) a non-linear DTI estimation procedure incorporating non-negative eigenvalue priors, and (iv) tract-specific region-of-interest (ROI) identification based on an objective geometry reference. Using these strategies in combination, radial diffusivity (λ⊥) was reproducibly measured in white matter (WM) tracts (adjusted mean [95% confidence interval]=0.25 [0.22, 0.29]µm2/ms), lower than previously reported λ⊥ values in the in vivo human spinal cord DTI literature. Radial diffusivity and fractional anisotropy (FA) measured in WM varied from rostral to caudal as did mean translational motion, likely reflecting respiratory motion effect. Given the considerable sensitivity of DTI measurements to motion artifact, we believe outlier detection is indispensable in spinal cord diffusion imaging. We also recommend using a mixed-effects model to account for systematic measurement bias depending on cord segment.
Directional diffusivity; Outlier rejection; Non-negative eigenvalue priors; Reduced FOV; Cardiac gating; Cervical spinal cord; Lateral corticospinal tract; Posterior column; Diffusion tensor imaging; Reproducibility
The authors have developed an automated computeraided diagnostic (CAD) scheme by using artificial neural networks (ANNs) on quantitative analysis of image data. Three separate ANNs were applied for detection of interstitial disease on digitized chest images. The first ANN was trained with horizontal profiles in regions of interest (ROIs) selected from normal and abnormal chest radiographs for distinguishing between normal and abnormal patterns. For training and testing of the second ANN, the vertical output patterns obtained from the 1st ANN were used for each ROI. The output value of the second ANN was used to distinguish between normal and abnormal ROIs with interstitial infiltrates. If the ratio of the number of abnormal ROIs to the total number of all ROIs in a chest image was greater than a specified threshold level, the image was classified as abnormal. In addition, the third ANN was applied to distinguish between normal and abnormal chest images. The combination of the rule-based method and the third ANN also was applied to the classification between normal and abnormal chest images. The performance of the ANNs was evaluated by means of receiver operating characteristic (ROC) analysis. The average Az value (area under the ROC curve) for distinguishing between normal and abnormal cases was 0.976±0.012 for 100 chest radiographs that were not used in training of ANNs. The results indicate that the ANN trained with image data can learn some statistical properties associated with interstitial infiltrates in chest radiographs.
interstitial infiltrate; computer-aided diagnosis; artificial neural network; chest radiograph
We propose an automated multi-atlas and multi-ROI based segmentation method for both skull-stripping of mouse brain and the ROI-labeling of mouse brain structures from the three dimensional (3D) magnetic resonance images (MRI). Three main steps are involved in our method. First, a region of interest (ROI) guided warping algorithm is designed to register multi-atlas images to the subject space, by considering more on the matching of image contents around the ROI boundaries which are more important for ROI labeling. Then, a multi-atlas and multi-ROI based deformable segmentation method is adopted to refine the ROI labeling result by deforming each ROI surface via boundary recognizers (i.e., SVM classifiers) trained on local surface patches. Finally, a local-mutual-information (MI) based multi-label fusion technique is proposed for allowing the atlases with better local image similarity with the subject to have more contributions in label fusion. The experimental results show that our method works better than the conventional methods on both in vitro and in vivo mouse brain datasets.
Mouse brain images; Segmentation; Multi-atlases; Multi-ROIs; Deformable segmentation; Label fusion
Exact timing is essential for functional MRI data analysis. Datasets are commonly measured using repeated 2D imaging methods, resulting in a temporal offset between slices. To compensate for this timing difference, slice-timing correction (i.e. temporal data interpolation) has been used as an fMRI pre-processing step for more than fifteen years. However, there has been an ongoing debate about the effectiveness and applicability of this method. This paper presents the first elaborated analysis of the impact of the slice-timing effect on simulated data for different fMRI paradigms and measurement parameters, taking into account data noise and smoothing effects. Here we show, depending on repetition time and paradigm design, slice-timing effects can significantly impair fMRI results and slice-timing correction methods can successfully compensate for these effects and therefore increase the robustness of the data analysis. In addition, our results from simulated data were supported by empirical in vivo datasets. Our findings suggest that slice-timing correction should be included in the fMRI pre-processing pipeline.
► Slice acquisition delays can degrade sensitivity of fMRI data analysis. ► Slice-timing correction during pre-processing suppresses estimator bias. ► Our findings based on extensive simulations are supported by in vivo data.
Functional MRI; Pre-processing; Analysis; Slice-timing correction
Studying connectivities among functional brain regions and the functional dynamics on brain networks has drawn increasing interest. A fundamental issue that affects functional connectivity and dynamics studies is how to determine the best possible functional brain regions or ROIs (regions of interest) for a group of individuals, since the connectivity measurements are heavily dependent on ROI locations. Essentially, identification of accurate, reliable and consistent corresponding ROIs is challenging due to the unclear boundaries between brain regions, variability across individuals, and nonlinearity of the ROIs. In response to these challenges, this paper presents a novel methodology to computationally optimize ROIs locations derived from task-based fMRI data for individuals so that the optimized ROIs are more consistent, reproducible and predictable across brains. Our computational strategy is to formulate the individual ROI location optimization as a group variance minimization problem, in which group-wise consistencies in functional/structural connectivity patterns and anatomic profiles are defined as optimization constraints. Our experimental results from multimodal fMRI and DTI data show that the optimized ROIs have significantly improved consistency in structural and functional profiles across individuals. These improved functional ROIs with better consistency could contribute to further study of functional interaction and dynamics in the human brain.
ROI optimization; structural connectivity; functional connectivity