Theoretical accuracy of model-based shape matching for measuring natural knee kinematics with single-plane fluoroscopy

Quantification of knee motion under dynamic, in vivo loaded conditions is necessary to understand how knee kinematics influence joint injury, disease, and rehabilitation. Though recent studies have measured three-dimensional knee kinematics by matching geometric bone models to single-plane fluoroscopic images, factors limiting the accuracy of this approach have not been thoroughly investigated. This study used a three-step computational approach to evaluate theoretical accuracy limitations due to the shape matching process alone. First, cortical bone models of the femur tibia/fibula, and patella were created from CT data. Next, synthetic (i.e., computer generated) fluoroscopic images were created by ray tracing the bone models in known poses. Finally, an automated matching algorithm utilizing edge detection methods was developed to align flat-shaded bone models to the synthetic images. Accuracy of the recovered pose parameters was assessed in terms of measurement bias and precision. Under these ideal conditions where other sources of error were eliminated, tibiofemoral poses were within 2 mm for sagittal plane translations and 1.5 deg for all rotations while patellofemoral poses were within 2 mm and 3 deg. However, statistically significant bias was found in most relative pose parameters. Bias disappeared and precision improved by a factor of two when the synthetic images were regenerated using flat shading (i.e., sharp bone edges) instead of ray tracing (i.e., attenuated bone edges). Analysis of absolute pose parameter errors revealed that the automated matching algorithm systematically pushed the flat-shaded bone models too far into the image plane to match the attenuated edges of the synthetic ray-traced images. These results suggest that biased edge detection is the primary factor limiting the theoretical accuracy of this single-plane shape matching procedure.     

A fast reconstruction algorithm for electron microscope tomography

We have implemented a Fast Fourier Summation algorithm for tomographic reconstruction of three-dimensional biological data sets obtained via transmission electron microscopy. We designed the fast algorithm to reproduce results obtained by the direct summation algorithm (also known as filtered or R-weighted backprojection). For two-dimensional images, the new algorithm scales as O(N(theta)M log M)+O(MN log N) operations, where N(theta) is the number of projection angles and M x N is the size of the reconstructed image. Three-dimensional reconstructions are constructed from sequences of two-dimensional reconstructions. We demonstrate the algorithm on real data sets. For typical sizes of data sets, the new algorithm is 1.5-2.5 times faster than using direct summation in the space domain. The speed advantage is even greater as the size of the data sets grows. The new algorithm allows us to use higher order spline interpolation of the data without additional computational cost. The algorithm has been incorporated into a commonly used package for tomographic reconstruction.     

Comparing genomes with duplications: a computational complexity point of view

In this paper, we are interested in the computational complexity of computing (dis)similarity measures between two genomes when they contain duplicated genes or genomic markers, a problem that happens frequently when comparing whole nuclear genomes. Recently, several methods ( [1], [2]) have been proposed that are based on two steps to compute a given (dis)similarity measure M between two genomes G_1 and G_2: first, one establishes a oneto- one correspondence between genes of G_1 and genes of G_2 ; second, once this correspondence is established, it defines explicitly a permutation and it is then possible to quantify their similarity using classical measures defined for permutations, like the number of breakpoints. Hence these methods rely on two elements: a way to establish a one-to-one correspondence between genes of a pair of genomes, and a (dis)similarity measure for permutations. The problem is then, given a (dis)similarity measure for permutations, to compute a correspondence that defines an optimal permutation for this measure. We are interested here in two models to compute a one-to-one correspondence: the exemplar model, where all but one copy are deleted in both genomes for each gene family, and the matching model, that computes a maximal correspondence for each gene family. We show that for these two models, and for three (dis)similarity measures on permutations, namely the number of common intervals, the maximum adjacency disruption (MAD) number and the summed adjacency disruption (SAD) number, the problem of computing an optimal correspondence is NP-complete, and even APXhard for the MAD number and SAD number.     

Modern three-dimensional conformal craniospinal radiotherapy

The main problem of craniospinal irradiation (CSI) is the matching of the fields. The use of a suitable technique is very important because matching of the fields is necessary to use for the optimal cancer irradiation of the long planning target volume (PTV). Since 2007, 8 patients have received CT-based, 3D-planned conformal CSI in our Institute. Patient immobilization was made in prone position in a vacuum bed, using skull and pelvis masks. Organ-at-risk (OAR) contours were made by radiographers. PTV was contoured by radiation oncologists. The prescribed dose to the PTV was 36 Gy with 1.8 Gy dose per fraction. In the planning process the following aspects were taken under consideration: all points of the PTV had to receive at least 95% of the prescribed dose (according to ICRU 50, 62); at junction field edges the overlapping parts were eliminated using a multisegmental technique, where the adjacent segment ends of the neighbouring fields were shifted two times 2 cm, so that the three equally weighted segments used in one field had 2-2 cm distance from each other. In the CSI planning the shape of the patient and so the length of the PTV has made a big emphasis on determining the number of field matching. Thus in some cases instead of two, only one field matching was sufficient - this could be achieved by increasing the source-to-skin distance (SSD) of the fields. The verification made with a solid-water phantom justified the precision of the field matching. The offset used at junction field edges in between one treatment facilitates the verification of field matching - and so the patient positioning. Thus the possibility of having overdosed regions could be reduced, which was very important from a radiation biological point of view.     

Development of separable electron spin resonance-computed tomography imaging for multiple radical species: an application to .OH and .NO

A method of separable ESR-CT (electron spin resonance-computed tomography) imaging for multiple radical species was developed and applied to imaging of .OH and .NO. The algorithm was improved by combining filtered back-projection with a modified algebraic reconstruction technique to enhance accuracy and shorten calculation time. With this algorithm, spectral-spatial images of the phantom consisting of 3-carbamoyl-2,2,5,5,-tetramethylpyrrolidine-N-oxyl and 2-phenyl-4,4,5,5,-tetramethylimidazoline-3-oxide-1-oxyl could be obtained in different directions by rotating the spatial axis. The spatial function of individual radicals was extracted by each of the two methods from each spectral-spatial image. The separative 2D images of each radical were individually constructed using the spatial function obtained with the two methods. By comparing the separative images with the phantom sample, the algorithm for separable ESR-CT imaging was established. This ESR-CT technique was combined with L-band ESR spectroscopy and applied to the separative imaging of .OH and .NO, which were spin trapped with 5, 5-dimethyl-1-pyrroline-N-oxide (DMPO) and Fe(2+)-N-methyl-D-glucamine dithiocarbamate complex, respectively. The ESR signal of DMPO-OH decreased gradually during data acquisition, and the decrease was calibrated by extrapolating the signal intensity to the beginning of data sampling. Both the position and size of the individual images for .OH and .NO were in very good agreement with the findings for the sample.     

A Local Agreement Filtering Algorithm for Transmission EM Reconstructions

We present LAFTER, an algorithm for de-noising single particle reconstructions from cryo-EM.Single particle analysis entails the reconstruction of high-resolution volumes from tens of thousands of particle images with low individual signal-to-noise. Imperfections in this process result in substantial variations in the local signal-to-noise ratio within the resulting reconstruction, complicating the interpretation of molecular structure. An effective local de-noising filter could therefore improve interpretability and maximise the amount of useful information obtained from cryo-EM maps.LAFTER is a local de-noising algorithm based on a pair of serial real-space filters. It compares independent half-set reconstructions to identify and retain shared features that have power greater than the noise. It is capable of recovering features across a wide range of signal-to-noise ratios, and we demonstrate recovery of the strongest features at Fourier shell correlation (FSC) values as low as 0.144 over a 256³-voxel cube. A fast and computationally efficient implementation of LAFTER is freely available.We also propose a new way to evaluate the effectiveness of real-space filters for noise suppression, based on the correspondence between two FSC curves: 1) the FSC between the filtered and unfiltered volumes, and 2) C_ref, the FSC between the unfiltered volume and a hypothetical noiseless volume, which can readily be estimated from the FSC between two half-set reconstructions.     

Activity-dependent regulation of synaptic size in Drosophila neuromuscular junctions

One of the fundamental questions in neural development is how neurons form synapses of the appropriate size for the efficient transfer of information across neural circuits. Here we investigated the mechanisms that bring about the size correlation between synapses and postsynaptic cells during development of Drosophila neuromuscular junctions (NMJs). To do this, we made use of a unique system in which two neighboring muscles (M6 and M7) are innervated by the same neurons. In mature NMJs, synaptic size on M6 is normally larger than that on M7, in accordance with the difference in muscle volume; this ensures the same extent of contraction of both muscles, and we refer to this correspondence as "matching". We found that matching was apparent in larvae 8 h after hatching, but not in newly hatched larvae despite the difference in muscle volume. When sensory inputs were suppressed by the expression of tetanus toxin in sensory neurons, matching did not occur, although synapses were able to grow. Matching was also suppressed by the inhibition of motoneuronal activity. These results suggest that matching is induced by regulating the rate of synaptic growth on M6 and M7 in an experience- and activity-dependent manner. It seems most likely that retrograde signals from the postsynaptic to the presynaptic cell convey the information about muscle cell size. We thus examined whether a candidate of retrograde signaling in NMJs, BMP signaling, is involved in matching. However, there was no effect on matching in BMP type II receptor gene mutants, suggesting that other experience-driven mechanisms besides BMP signaling are involved in the proper development of synapses.     

Comparison of photo‐matching algorithms commonly used for photographic capture–recapture studies

Photographic capture–recapture is a valuable tool for obtaining demographic information on wildlife populations due to its noninvasive nature and cost‐effectiveness. Recently, several computer‐aided photo‐matching algorithms have been developed to more efficiently match images of unique individuals in databases with thousands of images. However, the identification accuracy of these algorithms can severely bias estimates of vital rates and population size. Therefore, it is important to understand the performance and limitations of state‐of‐the‐art photo‐matching algorithms prior to implementation in capture–recapture studies involving possibly thousands of images. Here, we compared the performance of four photo‐matching algorithms; Wild‐ID, I3S Pattern+, APHIS, and AmphIdent using multiple amphibian databases of varying image quality. We measured the performance of each algorithm and evaluated the performance in relation to database size and the number of matching images in the database. We found that algorithm performance differed greatly by algorithm and image database, with recognition rates ranging from 100% to 22.6% when limiting the review to the 10 highest ranking images. We found that recognition rate degraded marginally with increased database size and could be improved considerably with a higher number of matching images in the database. In our study, the pixel‐based algorithm of AmphIdent exhibited superior recognition rates compared to the other approaches. We recommend carefully evaluating algorithm performance prior to using it to match a complete database. By choosing a suitable matching algorithm, databases of sizes that are unfeasible to match “by eye” can be easily translated to accurate individual capture histories necessary for robust demographic estimates.     

High resolution segmentation of cervical cells.

A major problem in the automation of cervical cytology screening is the segmentation of cell images. This paper presents the present status of the work on that problem at the University of Uppsala. A dual resolution system is used. Suspect malignant cells are located at 4 mu resolution. Each such cell is rescanned at 0.5 mu resolution at two different wavelengths, 530 and 570 nm. The nucleus and the cytoplasm are isolated each by two independent methods. For the nucleus adaptive thresholding in the histogram of the 570 nm image and a contouring in a radially transformed version of that image is used. For the cytoplasm a two dimensional thresholding in the 2D histogram and a contouring in a radially transformed version of the 530 nm image is used. If the two nuclear masks agree the surrounding area is checked for disturbing objects. If also the cytoplasm masks agree and are without disturbing objects the whole cell is accepted. The result of the cytoplasm masks agree and are without disturbing objects the whole cell is accepted. The result of the segmentation is thus three categories; free cells, free nuclei and rejected objects. The shape of the objects belonging to the former two categories is checked and irregularly shaped ones are rejected as probably consisting of several overlapping nuclei. Cells passing also this test are classified as normal or malignant. The experience from using this algorithm is discussed and areas for further research are pointed out.     

Geometric algorithms for the analysis of 2D-electrophoresis gels.

In proteomics, two-dimensional gel electrophoresis (2-DE) is a separation technique for proteins. The resulting protein spots can be identified either by using picking robots and subsequent mass spectrometry or by visual cross inspection of a new gel image with an already analyzed master gel. Difficulties especially arise from inherent noise and irregular geometric distortions in 2-DE images. Aiming at the automated analysis of large series of 2-DE images, or at the even more difficult interlaboratory gel comparisons, the bottleneck is to solve the two most basic algorithmic problems with high quality: Identifying protein spots and computing a matching between two images. For the development of the analysis software CAROl at Freie Universit?t Berlin, we have reconsidered these two problems and obtained new solutions which rely on methods from computational geometry. Their novelties are: 1. Spot detection is also possible for complex regions formed by several "merged" (usually saturated) spots; 2. User-defined landmarks are not necessary for the matching. Furthermore, images for comparison are allowed to represent different parts of the entire protein pattern, which only partially "overlap." The implementation is done in a client server architecture to allow queries via the internet. We also discuss and point at related theoretical questions in computational geometry.     

Manta Matcher: automated photographic identification of manta rays using keypoint features

For species which bear unique markings, such as natural spot patterning, field work has become increasingly more reliant on visual identification to recognize and catalog particular specimens or to monitor individuals within populations. While many species of interest exhibit characteristic markings that in principle allow individuals to be identified from photographs, scientists are often faced with the task of matching observations against databases of hundreds or thousands of images. We present a novel technique for automated identification of manta rays (Manta alfredi and Manta birostris) by means of a pattern‐matching algorithm applied to images of their ventral surface area. Automated visual identification has recently been developed for several species. However, such methods are typically limited to animals that can be photographed above water, or whose markings exhibit high contrast and appear in regular constellations. While manta rays bear natural patterning across their ventral surface, these patterns vary greatly in their size, shape, contrast, and spatial distribution. Our method is the first to have proven successful at achieving high matching accuracies on a large corpus of manta ray images taken under challenging underwater conditions. Our method is based on automated extraction and matching of keypoint features using the Scale‐Invariant Feature Transform (SIFT) algorithm. In order to cope with the considerable variation in quality of underwater photographs, we also incorporate preprocessing and image enhancement steps. Furthermore, we use a novel pattern‐matching approach that results in better accuracy than the standard SIFT approach and other alternative methods. We present quantitative evaluation results on a data set of 720 images of manta rays taken under widely different conditions. We describe a novel automated pattern representation and matching method that can be used to identify individual manta rays from photographs. The method has been incorporated into a website (mantamatcher.org) which will serve as a global resource for ecological and conservation research. It will allow researchers to manage and track sightings data to establish important life‐history parameters as well as determine other ecological data such as abundance, range, movement patterns, and structure of manta ray populations across the world.     

On the extraction and alignment of image edges

In this paper the types of image filtering, segmentation and correlational processes which are consistent with the ability of human observers to align edges in arbitrary two-dimensional images are considered. Edge-only versions of a number of images were computed by pre-filtering, segmenting and using laplacians to extract luminance boundaries. The matching of such edge-only images with the originals seems well predicted by a (derivative) matched filtering (cross-correlation) process showing some bias toward edges contained within the middle frequency range of the images used.     

Tissue counter analysis of histologic sections of melanoma: influence of mask size and shape, feature selection, statistical methods and tissue preparation.

BACKGROUND: Tissue counter analysis is an image analysis tool designed for the detection of structures in complex images at the macroscopic or microscopic scale. As a basic principle, small square or circular measuring masks are randomly placed across the image and image analysis parameters are obtained for each mask. Based on learning sets, statistical classification procedures are generated which facilitate an automated classification of new data sets. OBJECTIVE: To evaluate the influence of the size and shape of the measuring masks as well as the importance of feature selection, statistical procedures and technical preparation of slides on the performance of tissue counter analysis in microscopic images. As main quality measure of the final classification procedure, the percentage of elements that were correctly classified was used. STUDY DESIGN: HE-stained slides of 25 primary cutaneous melanomas were evaluated by tissue counter analysis for the recognition of melanoma elements (section area occupied by tumour cells) in contrast to other tissue elements and background elements. Circular and square measuring masks, various subsets of image analysis features and classification and regression trees compared with linear discriminant analysis as statistical alternatives were used. The percentage of elements that were correctly classified by the various classification procedures was assessed. In order to evaluate the applicability to slides obtained from different laboratories, the best procedure was automatically applied in a test set of another 50 cases of primary melanoma derived from the same laboratory as the learning set and two test sets of 20 cases each derived from two different laboratories, and the measurements of melanoma area in these cases were compared with conventional assessment of vertical tumour thickness. RESULTS: Square measuring masks were slightly superior to circular masks, and larger masks (64 or 128 pixels in diameter) were superior to smaller masks (8 to 32 pixels in diameter). As far as the subsets of image analysis features were concerned, colour features were superior to densitometric and Haralick texture features. Statistical moments of the grey level distribution were of least significance. CART (classification and regression tree) analysis turned out to be superior to linear discriminant analysis. In the best setting, 95% of melanoma tissue elements were correctly recognized. Automated measurement of melanoma area in the independent test sets yielded a correlation of r=0.846 with vertical tumour thickness (p<0.001), similar to the relationship reported for manual measurements. The test sets obtained from different laboratories yielded comparable results. CONCLUSIONS: Large, square measuring masks, colour features and CART analysis provide a useful setting for the automated measurement of melanoma tissue in tissue counter analysis, which can also be used for slides derived from different laboratories.     

Activity‐dependent regulation of synaptic size in Drosophila neuromuscular junctions

One of the fundamental questions in neural development is how neurons form synapses of the appropriate size for the efficient transfer of information across neural circuits. Here we investigated the mechanisms that bring about the size correlation between synapses and postsynaptic cells during development of Drosophila neuromuscular junctions (NMJs). To do this, we made use of a unique system in which two neighboring muscles (M6 and M7) are innervated by the same neurons. In mature NMJs, synaptic size on M6 is normally larger than that on M7, in accordance with the difference in muscle volume; this ensures the same extent of contraction of both muscles, and we refer to this correspondence as “matching”. We found that matching was apparent in larvae 8 h after hatching, but not in newly hatched larvae despite the difference in muscle volume. When sensory inputs were suppressed by the expression of tetanus toxin in sensory neurons, matching did not occur, although synapses were able to grow. Matching was also suppressed by the inhibition of motoneuronal activity. These results suggest that matching is induced by regulating the rate of synaptic growth on M6 and M7 in an experience‐ and activity‐dependent manner. It seems most likely that retrograde signals from the postsynaptic to the presynaptic cell convey the information about muscle cell size. We thus examined whether a candidate of retrograde signaling in NMJs, BMP signaling, is involved inmatching. However, there was no effect on matching inBMP type II receptor gene mutants, suggesting thatother experience‐driven mechanisms besides BMP signaling are involved in the proper development of synapses. © 2006 Wiley Periodicals, Inc. J Neurobiol, 2006     

Using weighted fixed neural networks for unsupervised fuzzy clustering

A novel algorithm for unsupervised fuzzy clustering is introduced. The algorithm uses a so-called Weighted Fixed Neural Network (WFNN) to store important and useful information about the topological relations in a given data set. The algorithm produces a weighted connected net, of weighted nodes connected by weighted edges, which reflects and preserves the topology of the input data set. The weights of the nodes and the edges in the resulting net are proportional to the local densities of data samples in input space. The connectedness of the net can be changed, and the higher the connectedness of the net is chosen, the fuzzier the system becomes. The new algorithm is computationally efficient when compared to other existing methods for clustering multi-dimensional data, such as color images.     

Contrast adaptation and contrast masking in human vision.

After a preliminary study of visual evoked potentials (VEPS) to a test grating seen in the presence of masks at different orientations, psychophysical data are presented showing the effects of adaptation and of masking on thresholds for detecting the same test grating. The test is a vertical grating of spatial frequency 2 cycles per degree; adapting and masking gratings differ from the test either in orientation or in spatial frequency. The effects of adaptation and masking are explained by a single mechanism model that assumes: (i) adaptation and masking both alter the contrast response (or transducer) function of the mechanism that detects the test; (ii) masks, but not adaptors, stimulate the mechanism that detects the test; and (iii) a test is detectable when it raises response level by a constant amount. The model incorporates two distinct tuning functions, a broad adaptive contrast function and a narrow effective contrast function. It accounts adequately for all the data, including the location and size of the facilitative dip found in some masking functions, the constant slopes of the threshold elevation segments of adaptation functions and the varying slopes of masking functions. It also predicts the sometimes surprising joint effects of adaptation followed by masking and of two masks operating simultaneously.     

Image Quality in Children with Low-Radiation Chest CT Using Adaptive Statistical Iterative Reconstruction and Model-Based Iterative Reconstruction

Objective

To evaluate noise reduction and image quality improvement in low-radiation dose chest CT images in children using adaptive statistical iterative reconstruction (ASIR) and a full model-based iterative reconstruction (MBIR) algorithm.

Methods

Forty-five children (age ranging from 28 days to 6 years, median of 1.8 years) who received low-dose chest CT scans were included. Age-dependent noise index (NI) was used for acquisition. Images were retrospectively reconstructed using three methods: MBIR, 60% of ASIR and 40% of conventional filtered back-projection (FBP), and FBP. The subjective quality of the images was independently evaluated by two radiologists. Objective noises in the left ventricle (LV), muscle, fat, descending aorta and lung field at the layer with the largest cross-section area of LV were measured, with the region of interest about one fourth to half of the area of descending aorta. Optimized signal-to-noise ratio (SNR) was calculated.

Result

In terms of subjective quality, MBIR images were significantly better than ASIR and FBP in image noise and visibility of tiny structures, but blurred edges were observed. In terms of objective noise, MBIR and ASIR reconstruction decreased the image noise by 55.2% and 31.8%, respectively, for LV compared with FBP. Similarly, MBIR and ASIR reconstruction increased the SNR by 124.0% and 46.2%, respectively, compared with FBP.

Conclusion

Compared with FBP and ASIR, overall image quality and noise reduction were significantly improved by MBIR. MBIR image could reconstruct eligible chest CT images in children with lower radiation dose.     

Shadows Alter Facial Expressions of Noh Masks

Background

A Noh mask, worn by expert actors during performance on the Japanese traditional Noh drama, conveys various emotional expressions despite its fixed physical properties. How does the mask change its expressions? Shadows change subtly during the actual Noh drama, which plays a key role in creating elusive artistic enchantment. We here describe evidence from two experiments regarding how attached shadows of the Noh masks influence the observers’ recognition of the emotional expressions.

Methodology/Principal Findings

In Experiment 1, neutral-faced Noh masks having the attached shadows of the happy/sad masks were recognized as bearing happy/sad expressions, respectively. This was true for all four types of masks each of which represented a character differing in sex and age, even though the original characteristics of the masks also greatly influenced the evaluation of emotions. Experiment 2 further revealed that frontal Noh mask images having shadows of upward/downward tilted masks were evaluated as sad/happy, respectively. This was consistent with outcomes from preceding studies using actually tilted Noh mask images.

Conclusions/Significance

Results from the two experiments concur that purely manipulating attached shadows of the different types of Noh masks significantly alters the emotion recognition. These findings go in line with the mysterious facial expressions observed in Western paintings, such as the elusive qualities of Mona Lisa’s smile. They also agree with the aesthetic principle of Japanese traditional art “yugen (profound grace and subtlety)”, which highly appreciates subtle emotional expressions in the darkness.     

Image quality assessment of an iterative reconstruction algorithm applied to abdominal CT imaging

PurposeTo compare the noise and accuracy on images of the whole porcine liver acquired with iterative reconstruction (IMR, Philips Healthcare, Cleveland, OH, USA) and filtered back projection (FBP) methods.Materials and methodsWe used non-enhanced porcine liver to simulate the human liver and acquired it 100 times to obtain the average FBP value as the ground-truth. The mean and the standard deviation (“inter-scan SD”) of the pixel values on the 100 image acquisitions were calculated for FBP and for three levels of IMR (L1, L2, and L3). We also calculated the noise power spectrum (NPS) and the normalized NPS for the 100 image acquisitions.ResultsThe spatial SD for the porcine liver parenchyma on these slices was 9.92, 4.37, 3.63, and 2.30 Hounsfield units with FBP, IMR-L1, IMR-L2, and IMR-L3, respectively. The detectability of small faint features was better on single IMR than single FBP images. The inter-scan SD value for IMR-L3 images was 53% larger at the liver edges than at the liver parenchyma; it was only 10% larger on FBP images. Assessment of the normalized NPS showed that the noise on IMR images was comprised primarily of low-frequency components.ConclusionIMR images yield the same structure informations as FBP images and image accuracy is maintained. On level 3 IMR images the image noise is more strongly suppressed than on IMR images of the other levels and on FBP images.     

Possibility Study of Scale Invariant Feature Transform (SIFT) Algorithm Application to Spine Magnetic Resonance Imaging

The purpose of this study is an application of scale invariant feature transform (SIFT) algorithm to stitch the cervical-thoracic-lumbar (C-T-L) spine magnetic resonance (MR) images to provide a view of the entire spine in a single image. All MR images were acquired with fast spin echo (FSE) pulse sequence using two MR scanners (1.5 T and 3.0 T). The stitching procedures for each part of spine MR image were performed and implemented on a graphic user interface (GUI) configuration. Moreover, the stitching process is performed in two categories; manual point-to-point (mPTP) selection that performed by user specified corresponding matching points, and automated point-to-point (aPTP) selection that performed by SIFT algorithm. The stitched images using SIFT algorithm showed fine registered results and quantitatively acquired values also indicated little errors compared with commercially mounted stitching algorithm in MRI systems. Our study presented a preliminary validation of the SIFT algorithm application to MRI spine images, and the results indicated that the proposed approach can be performed well for the improvement of diagnosis. We believe that our approach can be helpful for the clinical application and extension of other medical imaging modalities for image stitching.