We investigated gaze-contingent fusion of infrared imagery during visual search. Eye movements were monitored while subjects searched for and identified human targets in images captured simultaneously in the short-wave (SWIR) and long-wave (LWIR) infrared bands. Based on the subject’s gaze position, the search display was updated such that imagery from one sensor was continuously presented to the subject’s central visual field (“center”) and another sensor was presented to the subject’s non-central visual field (“surround”). Analysis of performance data indicated that, compared to the other combinations, the scheme featuring SWIR imagery in the center region and LWIR imagery in the surround region constituted an optimal combination of the SWIR and LWIR information: it inherited the superior target detection performance of LWIR imagery and the superior target identification performance of SWIR imagery. This demonstrates a novel method for efficiently combining imagery from two infrared sources as an alternative to conventional image fusion.

Proximity cues including perspective, motion parallax, etc., are important depth cues besides disparity cue and blur (defocus) cue. Disparity and blur are now known to have important effects on visual comfort when viewing stereoscopic content. However, the effects of proximity cues are still unknown. In order to explore this question, the authors conducted two experiments. In the first one, the instant effects of proximity cues on the visual comfort of 3D stimuli were assessed using relative comparison. There were, in total, three sessions in this experiment, in which all 3D stimuli were compared with different baselines. The results showed that proximity cues also have significant effects on visual comfort when viewing stereoscopic content. Such effects do not vary significantly with disparity cue, blur cue and the selection of baseline stimuli. In the second experiment, the authors further tested the stimuli with proximity cues, and the results were in accordance with many existing literature results when the stimuli were viewed consecutively. Their results suggest that the perspective information and radial optic flow in the proximity cues may increase vergence–accommodation conflict and therefore reduce visual comfort. However, the effects seem to be limited during consecutive viewing.

Light field 3D displays represent a major step forward in visual realism, providing glasses-free spatial vision of real or virtual scenes. Applications that capture and process live imagery have to process data captured by potentially tens to hundreds of cameras and control tens to hundreds of projection engines making up the human perceivable 3D light field using a distributed processing system. The associated massive data processing is difficult to scale beyond a specific number and resolution of images, limited by the capabilities of the individual computing nodes. The authors therefore analyze the bottlenecks and data flow of the light field conversion process and identify possibilities to introduce better scalability. Based on this analysis they propose two different architectures for distributed light field processing. To avoid using uncompressed video data all along the processing chain, the authors also analyze how the operation of the proposed architectures can be supported by existing image/video codecs.

The authors present a new method of writer identification, employing the full power of multiple experiments, which yields a statistically significant result. Each individual binarized and segmented character is represented as a histogram of 512 binary pixel patterns—3 × 3 black and white patches. In the process of comparing two given inscriptions under a “single author” assumption, the algorithm performs a Kolmogorov–Smirnov test for each letter and each patch. The resulting p-values are combined using Fisher’s method, producing a single p-value. Experiments on both Modern and Ancient Hebrew data sets demonstrate the excellent performance and robustness of this approach.

Model-based image reconstruction (MBIR) techniques have the potential to generate high quality images from noisy measurements and a small number of projections which can reduce the x-ray dose in patients. These MBIR techniques rely on projection and backprojection to refine an image estimate. One of the widely used projectors for these modern MBIR based technique is called branchless distance driven (DD) projection and backprojection. While this method produces superior quality images, the computational cost of iterative updates keeps it from being ubiquitous in clinical applications. In this paper, we provide several new parallelization ideas for concurrent execution of the DD projectors in multi-GPU systems using CUDA programming tools. We have introduced some novel schemes for dividing the projection data and image voxels over multiple GPUs to avoid runtime overhead and inter-device synchronization issues. We have also reduced the complexity of overlap calculation of the algorithm by eliminating the common projection plane and directly projecting the detector boundaries onto image voxel boundaries. To reduce the time required for calculating the overlap between the detector edges and image voxel boundaries, we have proposed a pre-accumulation technique to accumulate image intensities in perpendicular 2D image slabs (from a 3D image) before projection and after backprojection to ensure our DD kernels run faster in parallel GPU threads. For the implementation of our iterative MBIR technique we use a parallel multi-GPU version of the alternating minimization (AM) algorithm with penalized likelihood update. The time performance using our proposed reconstruction method with Siemens Sensation 16 patient scan data shows an average of 24 times speedup using a single TITAN X GPU and 74 times speedup using 3 TITAN X GPUs in parallel for combined projection and backprojection.

Projection-based augmented reality systems overlay digital information directly on real objects, while at the same time use cameras to capture the scene information. A common problem with such systems is that cameras see the projected image besides the real objects to some degree. This crosstalk reduces the object detection and digital content registration abilities. The authors propose a novel time sharing-based technique that facilitates the real and digital content decoupling in real time without crosstalk. The proposed technique is based on time sequential operation between a MEMS scanner-based mobile projector and rolling shutter image sensor. A MEMS mirror- based projector scans light beam in raster pattern pixel by pixel and completes full frame projection over a refresh period, while a rolling shutter image sensor sequentially collects scene light row by row. In the proposed technique, the image sensor is synchronized with scanning MEMS mirror and precisely follows the display scanner with a half-period lag to make the displayed content completely invisible for camera. An experimental setup consisting of laser pico projector, an image sensor, and a delay and amplifier circuit is developed. The performance of proposed technique is evaluated by measuring the crosstalk in captured content and sensor exposure limit. The results show 0% crosstalk in captured content up to 8 ms sensor exposure. High capture frame rate (up to 45 fps) is achieved by cyclically triggering a 3.2 MP, 60 fps CMOS sensor and using a 60 Hz pico projector.