Invariant image classification using triple-correlation-basedneural networks

Triple-correlation-based neural networks are introduced and used in this paper for invariant classification of 2D gray scale images. Third-order correlations of an image are appropriately clustered, in spatial or spectral domain, to generate an equivalent image representation that is invariant with respect to translation, rotation, and dilation. An efficient implementation scheme is also proposed, which is robust to distortions, insensitive to additive noise, and classifies the original image using adequate neural network architectures applied directly to 2D image representations. Third-order neural networks are shown to be a specific category of triple-correlation-based networks, applied either to binary or gray-scale images. A simulation study is given, which illustrates the theoretical developments, using synthetic and real image data.     

Relevance of pharmacokinetics and pharmacodynamics

This ex post facto study investigated the relationship of social support from two workplace sources, unit managers and registered nurse (RNs) coworkers, to work-related stress among staff nurses. Sample participants included 51 full-time RNs from a variety of clinical specialties in one nursing organization. The three theoretical components--affect, affirmation, and aid--were measured by the Norbeck Social Support Questionnaire. The Nursing Stress Scale was used to quantify the levels of perceived work-related stress. Statistically significant negative correlations were found between affect from unit managers (r = -299; P = 0.15) and aid from coworkers (r = -.294; P = .036) on levels of perceived work-related stress. Additionally, the increased age of the subjects (r = -.342; P = .007), years of nursing experience (r = -252; P = .037), and length of employment (r = -329; P = .009) had a negative relationship with work stress. The results of this research demonstrated that two components of social support--affect (emotional support) from unit managers and aid (services, information, materials) from coworkers--were linked to lower stress scores in this sample of staff nurses.     

A non-specific lipid transfer protein from Arabidopsis is a cell wall protein

Lipid transfer proteins (LTPs), mediate the transfer of phospholipids between membranes in vitro. However, the in vivo function of LTPs is not known. To determine the precise location of a non-specific LTP from Arabidopsis, a cDNA clone was used to produce an Arabidopsis LTP:protein A fusion. Antibodies raised against the fusion were used to localize the Arabidopsis LTP by immunoelectron microscopy. LTP was found to be located in the cell wall, mainly in epidermal cells. This location appears to be inconsistent with the proposed role of the protein in intracellular lipid transfer.     

The influence of trigger configurations and laminate lay-up on the failure mode of composite crush cylinders

In this paper the results are presented of a test program on the energy absorption of composite cylinders loaded in compression. The influence of the laminate lay-up and of the trigger configuration were determined. Two different failure modes for the different laminates and triggers were observed: a splaying mode and a fragmentation mode. The splaying mode is more efficient in absorbing energy. The failure mode did not change during the crushing process.     

Cable insulation density profile determined from ultrasonicvelocity measurements

Density results obtained on the same cable sample with a column and from ultrasonic measurements are reported. A calibration curve relating ultrasonic velocity and cable insulation density is given for PE and crosslinked PE (XLPE) cables. The limits and advantages of deducing density profiles of cable insulation using a focused, ultrasound beam rather than the more time-consuming column measurements are discussed     

Magnetic resonance imaging study on the results of surgery for cervical compression myelopathy

The morphologic changes and signal intensity of the spinal cord on preoperative magnetic resonance images were correlated with postoperative outcomes in 74 patients undergoing decompressive cervical surgery for compressive myelopathy. The transverse area of the spinal cord on T1-weighted images at the level of maximum compression was closely correlated with the severity of myelopathy, duration of disease, and recovery rate as determined by the Japanese Orthopaedic Association score. In patients with ossification of the posterior longitudinal ligament or cervical spondylotic myelopathy, the increased intramedullary T2-weighted magnetic resonance imaging signal at the site of maximal cord compression and duration of disease significantly influenced the rate of recovery. A multiple regression equation was then developed with these three variables to predict surgical outcomes.