In this work we propose to combine a supervoxel-based image representation with the concept of graph cuts as an efficient optimization technique for 3D deformable image registration. Due to the pixels/voxels-wise graph construction, the use of graph cuts in this context has been mainly limited to 2D applications. However, our work overcomes some of the previous limitations by posing the problem on a graph created by adjacent supervoxels, where the number of nodes in the graph is reduced from the number of voxels to the number of supervoxels. We demonstrate how a supervoxel image representation, combined with graph cuts-based optimization can be applied to 3D data. We further show that the application of a relaxed graph representation of the image, followed by guided image filtering over the estimated deformation field, allows us to model 'sliding motion'. Applying this method to lung image registration, results in highly accurate image registration and anatomically plausible estimations of the deformations. Evaluation of our method on a publicly available Computed Tomography lung image dataset (www.dir-lab.com) leads to the observation that our new approach compares very favorably with state-of-the-art in continuous and discrete image registration methods achieving Target Registration Error of 1.16mm on average per landmark.

High-speed cameras explore more details than normal cameras in the time sequence, while the conventional video sampling suffers from the trade-off between temporal and spatial resolutions due to the sensor's physical limitation. Compressive sensing overcomes this obstacle by combining the sampling and compression procedures together. A single-pixel-based real-time video acquisition is proposed to record dynamic scenes, and a fast nonconvex algorithm for the nonconvex sorted ℓ regularization is applied to reconstruct frame differences using few numbers of measurements. Then, an edge-detection-based denoising method is employed to reduce the error in the frame difference image. The experimental results show that the proposed algorithm together with the single-pixel imaging system makes compressive video cameras available.

Visual prostheses require an effective representation method due to the limited display condition which has only 2 or 3 levels of grayscale in low resolution. Edges derived from abrupt luminance changes in images carry essential information for object recognition. Typical binary (black and white) edge images have been used to represent features to convey essential information. However, in scenes with a complex cluttered background, the recognition rate of the binary edge images by human observers is limited and additional information is required. The polarity of edges and cusps (black or white features on a gray background) carries important additional information; the polarity may provide shape from shading information missing in the binary edge image. This depth information may be restored by using bipolar edges. We compared object recognition rates from 16 binary edge images and bipolar edge images by 26 subjects to determine the possible impact of bipolar filtering in visual prostheses with 3 or more levels of grayscale. Recognition rates were higher with bipolar edge images and the improvement was significant in scenes with complex backgrounds. The results also suggest that erroneous shape from shading interpretation of bipolar edges resulting from pigment rather than boundaries of shape may confound the recognition.

We investigate the use of different trabecular bone descriptors and advanced machine learning tech niques to complement standard bone mineral density (BMD) measures derived from dual-energy x-ray absorptiometry (DXA) for improving clinical assessment of osteoporotic fracture risk. For this purpose, volumes of interest were extracted from the head, neck, and trochanter of 146 proximal femur specimens on multidetector computer tomography. The trabecular bone captured was characterized with (1) statistical moments of the BMD distribution, (2) geometrical features derived from the scaling index method (SIM), and (3) morphometric parameters, such as bone fraction, trabecular thickness, etc. Feature sets comprising DXA BMD and such supplemental features were used to predict the failure load (FL) of the specimens, previously determined through biomechanical testing, with multiregression and support vector regression. Prediction performance was measured by the root mean square error (RMSE); correlation with measured FL was evaluated using the coefficient of determination . The best prediction performance was achieved by a combination of DXA BMD and SIM-derived geometric features derived from the femoral head (RMSE: 0.869 ± 0.121, : 0.68 ± 0.079), which was significantly better than DXA BMD alone (RMSE: 0.948 ± 0.119, : 0.61 ± 0.101) ( < 10). For multivariate feature sets, SVR outperformed multiregression ( < 0.05). These results suggest that supplementing standard DXA BMD measurements with sophisticated femoral trabecular bone characterization and supervised learning techniques can significantly improve biomechanical strength prediction in proximal femur specimens.

The usual tomography is achieved by acquiring measurements around an object with multiple angles. The possibility of obtaining a fluorescence tomographic image from measurements at only one angle is explored. Instead of rotating around the object, the camera (or the objective lens) moves toward (or away from) the object and takes photographs while the camera's focal plane passes through the object. The volume of stacked two-dimensional pictures forms a blurred three-dimensional image. The true image can be obtained by deconvolving the system's point spread function. Simplified computer simulations are carried out to verify the feasibility of the proposed method. The computer simulations indicate that it is feasible to obtain a tomographic image by using the in-and-out motion to acquire data.

The Katsevich algorithm is a breakthrough in the theoretically exact algorithms for helical cone beam CT. For future application in medical and industrial CT, determining how to implement it efficiently and accurately is the main task. We analyzed the slope law and intersection law of the k-lines, finding that the k-lines are not intersecting if the half maximal fan angle (HMFA) is less than 21° (numerical solution, so it is approximate) and that the helical pitch and HMFA determine the depth of parallelism of k-lines. Using an appropriate pitch and a HWFA that is less than 21°, one can use a simplified Katsevich algorithm, whose filtration process can be done on the rows of the detector panel so that the pre-weighting, pre-rebinning, post-rebinning and post-weighting steps are all canceled. Simulation experiments show that the simplified algorithm can obtain highly precise images at a faster speed. Our results are intended to be valuable to those who are working on efficient implementations of the Katsevich-type algorithms.

We discuss improved image reconstruction and segmentation in a framework we term model-controlled flooding (MCF). This extends the watershed transform for segmentation by allowing the integration of a priori information about image objects into flooding simulation processes. Modeling the initial seeding, region growing, and stopping rules of the watershed flooding process allows users to customize the simulation with user-defined or default model functions incorporating prior information. It also extends a more general class of transforms based on connected attribute filters by allowing the modification of connected components of a grayscale image, thus providing more flexibility in image reconstruction. MCF reconstruction defines images with desirable features for further segmentation using existing methods and can lead to substantial improvements. We demonstrate the MCF framework using a size transform that extends grayscale area opening and attribute thickening/thinning, and give examples from several areas: concealed object detection, speckle counting in biological single cell studies, and analyses of benchmark microscopic image data sets. MCF achieves benchmark error rates well below those reported in the recent literature and in comparison with other algorithms, while being easily adapted to new imaging contexts.

In this paper we describe and evaluate a fast implementation of a classical block matching motion estimation algorithm for multiple Graphical Processing Units (GPUs) using the Compute Unified Device Architecture (CUDA) computing engine. The implemented block matching algorithm (BMA) uses summed absolute difference (SAD) error criterion and full grid search (FS) for finding optimal block displacement. In this evaluation we compared the execution time of a GPU and CPU implementation for images of various sizes, using integer and non-integer search grids.The results show that use of a GPU card can shorten computation time by a factor of 200 times for integer and 1000 times for a non-integer search grid. The additional speedup for non-integer search grid comes from the fact that GPU has built-in hardware for image interpolation. Further, when using multiple GPU cards, the presented evaluation shows the importance of the data splitting method across multiple cards, but an almost linear speedup with a number of cards is achievable.In addition we compared execution time of the proposed FS GPU implementation with two existing, highly optimized non-full grid search CPU based motion estimations methods, namely implementation of the Pyramidal Lucas Kanade Optical flow algorithm in OpenCV and Simplified Unsymmetrical multi-Hexagon search in H.264/AVC standard. In these comparisons, FS GPU implementation still showed modest improvement even though the computational complexity of FS GPU implementation is substantially higher than non-FS CPU implementation.We also demonstrated that for an image sequence of 720×480 pixels in resolution, commonly used in video surveillance, the proposed GPU implementation is sufficiently fast for real-time motion estimation at 30 frames-per-second using two NVIDIA C1060 Tesla GPU cards.

The authors developed an integrated computer-aided detection (CAD) scheme for detecting and classifying metaphase chromosomes as well as assessing its performance and robustness. This scheme includes an automatic metaphase-finding module and a karyotyping module and it was applied to a testing database with 200 digital microscopic images. The automatic metaphase-finding module detects analyzable metaphase cells using a feature-based artificial neural network (ANN). The ANN-generated outputs are analyzed by a receiver operating characteristics (ROC) method and an area under the ROC curve is 0.966. Then, the automatic karyotyping module classifies individual chromosomes of this cell into 24 types. In this module, a two-layer decision tree-based classifier with eight ANNs established in its connection nodes was optimized by a genetic algorithm. Chromosomes are first classified into seven groups by the ANN in the first layer. The chromosomes in these groups are then separately classified by seven ANNs into 24 types in the second layer. The classification accuracy is 94.5% in the first layer. Six ANNs achieved the accuracy above 95% and only one had lessened performance (80.6%) in the second layer. The overall classification accuracy is 91.5% as compared to 86.7% in the previous study using two independent datasets randomly acquired from our genetic laboratory. The results demonstrate that our automated scheme achieves high and robust performance in identification and classification of metaphase chromosomes.

Confocal microscopes (CM) are routinely used for building 3-D images of microscopic structures. Nonideal imaging conditions in a white-light CM introduce additive noise and blur. The optical section images need to be restored prior to quantitative analysis. We present an adaptive noise filtering technique using Karhunen-Loéve expansion (KLE) by the method of snapshots and a ringing metric to quantify the ringing artifacts introduced in the images restored at various iterations of iterative Lucy-Richardson deconvolution algorithm. The KLE provides a set of basis functions that comprise the optimal linear basis for an ensemble of empirical observations. We show that most of the noise in the scene can be removed by reconstructing the images using the KLE basis vector with the largest eigenvalue. The prefiltering scheme presented is faster and does not require prior knowledge about image noise. Optical sections processed using the KLE prefilter can be restored using a simple inverse restoration algorithm; thus, the methodology is suitable for real-time image restoration applications. The KLE image prefilter outperforms the temporal-average prefilter in restoring CM optical sections. The ringing metric developed uses simple binary morphological operations to quantify the ringing artifacts and confirms with the visual observation of ringing artifacts in the restored images.

Tone mapping refers to the conversion of luminance values recorded by a digital camera or other acquisition device, to the luminance levels available from an output device, such as a monitor or a printer. Tone mapping can improve the appearance of rendered images. Although there are a variety of algorithms available, there is little information about the image tone characteristics that produce pleasing images. We devised an experiment where preferences for images with different tone characteristics were measured. The results indicate that there is a systematic relation between image tone characteristics and perceptual image quality for images containing faces. For these images, a mean face luminance level of 46-49 CIELAB L* units and a luminance standard deviation (taken over the whole image) of 18 CIELAB L* units produced the best renderings. This information is relevant for the design of tone-mapping algorithms, particularly as many images taken by digital camera users include faces.