Real-Time Multimodal Retinal Image Registration for a Computer Assisted Laser Photocoagulation System

An algorithm for the real-time registration of a retinal video sequence captured with a scanning digital ophthalmoscope (SDO) to a retinal composite image is presented. This method is designed for a computer-assisted retinal laser photocoagulation system to compensate for retinal motion and hence enhance the accuracy, speed, and patient safety of retinal laser treatments. The procedure combines intensity and feature-based registration techniques. For the registration of an individual frame, the translational frame-to-frame motion between preceding and current frame is detected by normalized cross correlation. Next, vessel points on the current video frame are identified and an initial transformation estimate is constructed from the calculated translation vector and the quadratic registration matrix of the previous frame. The vessel points are then iteratively matched to the segmented vessel centerline of the composite image to refine the initial transformation and register the video frame to the composite image. Criteria for image quality and algorithm convergence are introduced, which assess the exclusion of single frames from the registration process and enable a loss of tracking signal if necessary. The algorithm was successfully applied to ten different video sequences recorded from patients. It revealed an average accuracy of 2.47 ± 2.0 pixels (∼23.2 ± 18.8 μm) for 2764 evaluated video frames and demonstrated that it meets the clinical requirements.

An insulin infusion advisory system based on autotuning nonlinear model-predictive control

This paper aims at the development and evaluation of a personalized insulin infusion advisory system (IIAS), able to provide real-time estimations of the appropriate insulin infusion rate for type 1 diabetes mellitus (T1DM) patients using continuous glucose monitors and insulin pumps. The system is based on a nonlinear model-predictive controller (NMPC) that uses a personalized glucose-insulin metabolism model, consisting of two compartmental models and a recurrent neural network. The model takes as input patient's information regarding meal intake, glucose measurements, and insulin infusion rates, and provides glucose predictions. The predictions are fed to the NMPC, in order for the latter to estimate the optimum insulin infusion rates. An algorithm based on fuzzy logic has been developed for the on-line adaptation of the NMPC control parameters. The IIAS has been in silico evaluated using an appropriate simulation environment (UVa T1DM simulator). The IIAS was able to handle various meal profiles, fasting conditions, interpatient variability, intraday variation in physiological parameters, and errors in meal amount estimations.

A new system for computer-aided preoperative planning and intraoperative navigation during corrective jaw surgery

A new system for computer-aided corrective surgery of the jaws has been developed and introduced clinically. It combines three-dimensional (3-D) surgical planning with conventional dental occlusion planning. The developed software allows simulating the surgical correction on virtual 3-D models of the facial skeleton generated from computed tomography (CT) scans. Surgery planning and simulation include dynamic cephalometry, semi-automatic mirroring, interactive cutting of bone and segment repositioning. By coupling the software with a tracking system and with the help of a special registration procedure, we are able to acquire dental occlusion plans from plaster model mounts. Upon completion of the surgical plan, the setup is used to manufacture positioning splints for intraoperative guidance. The system provides further intraoperative assistance with the help of a display showing jaw positions and 3-D positioning guides updated in real time during the surgical procedure. The proposed approach offers the advantages of 3-D visualization and tracking technology without sacrificing long-proven cast-based techniques for dental occlusion evaluation. The system has been applied on one patient. Throughout this procedure, we have experienced improved assessment of pathology, increased precision, and augmented control.

Acute device-based blood pressure reduction: electrical activation of the carotid baroreflex in patients undergoing elective carotid surgery

Carotid sinus baroreceptors are involved in controlling blood pressure (BP) by providing input to the cardiovascular regulatory centers of the medulla. The acute effect of temporarily placing an electrode on the carotid sinus wall to electrically activate the baroreflex was investigated. We studied 11 patients undergoing elective carotid surgery. Baseline BP was 146+30/66+/-17 mm Hg and heart rate (HR) 72+/-7 bpm (mean +/- standard deviation). An electrode was placed upon the carotid sinus and after obtaining a steady state baseline of BP and HR, an electric current was applied and increased in 1-volt increments. A voltage dependent and highly significant reduction in BP was observed which averaged 18+/-26* and 8.0+/-12 mm Hg for systolic BP and diastolic BP, respectively. Maximal reductions occurred at 4.4+/-1.2 V: 23+/-24 mm Hg*, 16+/-10 mm Hg* and 7+/-12 bpm* for systolic BP, diastolic BP and HR, respectively ( = p <.05). Thus, electrical stimulation of the carotid sinus activates the carotid baroreflex resulting in a reduction in BP and HR. This presents a proof of concept for device based baroreflex modulation in acute BP regulation and adds to the available data which provide a rationale for evaluating this system in the context of chronic BP reduction in hypertensive patients.

An implantable carotid sinus baroreflex activating system: surgical technique and short-term outcome from a multi-center feasibility trial for the treatment of resistant hypertension

OBJECTIVES: To assess perioperative outcomes and blood pressure (BP) responses to an implantable carotid sinus baroreflex activating system being investigated for the treatment of resistant hypertension. METHODS: We report on the first seventeen patients enrolled in a multicenter study. Bilateral perivascular carotid sinus electrodes (CSL) and a pulse generator (IPG) are permanently implanted. Optimal placement of the CSL is determined by intraoperative BP responses to test activations. Acute BP responses were tested postoperatively and during the first four months of follow-up. RESULTS: Prior to implant, BP was 189.6+/-27.5/110.7+/-15.3 mmHg despite stable therapy (5.2+/-1.8 antihypertensive drugs). The mean procedure time was 202+/-43 minutes. No perioperative strokes or deaths occurred. System tests performed 1 or up to 3 days postoperatively resulted in significant (all p < or = 0.0001) mean maximum reduction, with standard deviations and 95% confidence limits for systolic BP, diastolic BP and heart rate of 28+/-22 (17, 39) mmHg, 16+/-11 (10, 22) mmHg and 8+/-4 (6, 11) BPM, respectively. Repeated testing during 3 months of therapeutic electrical activation demonstrated a durable response. CONCLUSIONS: These preliminary data suggest an acceptable safety of the procedure with a low rate of adverse events and support further clinical development of baroreflex activation as a new concept to treat resistant hypertension.

Effects of butorphanol on the withdrawal reflex using threshold, suprathreshold and repeated subthreshold electrical stimuli in conscious horses

OBJECTIVE: To assess the effects of a single intravenous dose of butorphanol (0.1 mg kg(-1)) on the nociceptive withdrawal reflex (NWR) using threshold, suprathreshold and repeated subthreshold electrical stimuli in conscious horses. STUDY DESIGN: 'Unblinded', prospective experimental study. ANIMALS: Ten adult horses, five geldings and five mares, mean body mass 517 kg (range 487-569 kg). METHODS: The NWR was elicited using single transcutaneous electrical stimulation of the palmar digital nerve. Repeated stimulations were applied to evoke temporal summation. Surface electromyography was performed to record and quantify the responses of the common digital extensor muscle to stimulation and behavioural reactions were scored. Before butorphanol administration and at fixed time points up to 2 hours after injection, baseline threshold intensities for NWR and temporal summation were defined and single suprathreshold stimulations applied. Friedman repeated-measures analysis of variance on ranks and Wilcoxon signed-rank test were used with the Student-Newman-Keul's method applied post-hoc. The level of significance (alpha) was set at 0.05. RESULTS: Butorphanol did not modify either the thresholds for NWR and temporal summation or the reaction scores, but the difference between suprathreshold and threshold reflex amplitudes was reduced when single stimulation was applied. Upon repeated stimulation after butorphanol administration, a significant decrease in the relative amplitude was calculated for both the 30-80 and the 80-200 millisecond intervals after each stimulus, and for the whole post-stimulation interval in the right thoracic limb. In the left thoracic limb a decrease in the relative amplitude was found only in the 30-80 millisecond epoch. CONCLUSION: Butorphanol at 0.1 mg kg(-1) has no direct action on spinal Adelta nociceptive activity but may have some supraspinal effects that reduce the gain of the nociceptive system. CLINICAL RELEVANCE: Butorphanol has minimal effect on sharp immediate Adelta-mediated pain but may alter spinal processing and decrease the delayed sensations of pain.

A computer-aided diagnostic system to characterize CT focal liver lesions: design and optimization of a neural network classifier

In this paper, a computer-aided diagnostic (CAD) system for the classification of hepatic lesions from computed tomography (CT) images is presented. Regions of interest (ROIs) taken from nonenhanced CT images of normal liver, hepatic cysts, hemangiomas, and hepatocellular carcinomas have been used as input to the system. The proposed system consists of two modules: the feature extraction and the classification modules. The feature extraction module calculates the average gray level and 48 texture characteristics, which are derived from the spatial gray-level co-occurrence matrices, obtained from the ROIs. The classifier module consists of three sequentially placed feed-forward neural networks (NNs). The first NN classifies into normal or pathological liver regions. The pathological liver regions are characterized by the second NN as cyst or "other disease." The third NN classifies "other disease" into hemangioma or hepatocellular carcinoma. Three feature selection techniques have been applied to each individual NN: the sequential forward selection, the sequential floating forward selection, and a genetic algorithm for feature selection. The comparative study of the above dimensionality reduction methods shows that genetic algorithms result in lower dimension feature vectors and improved classification performance.

Assessment of the wind-up phenomenon in the equine nociceptive trigeminal system.

Repeated sub-threshold nociceptive electrical stimulation resulting in temporal summation of the limb nociceptive withdrawal reflex is a well-established non-invasive model to investigate the wind-up phenomenon in horses. Due to structural similarities of the trigeminal sensory nucleus to the dorsal horn of the spinal cord, temporal summation should be evoked by repeated transcutaneous electrical stimulation of trigeminal afferents. To evaluate this hypothesis repeated transcutaneous electrical stimulation was applied to the supraorbital and infraorbital nerves of 10 horses. Stimulation intensities varied between 0.5 and 1.3 times the trigemino-cervical reflex threshold defined for single stimulation. Evoked electromyographic activity of the orbicularis oculi, splenius and cleidomastoideus muscles was recorded and the signals analysed in the previously established epochs typical to the early and late component of the blink reflex and to the trigemino-cervical reflex. Behavioural reactions were evaluated with the aid of numerical rating scale. The nociceptive late component and the trigemino-cervical reflex were not elicited by sub-threshold intensity repeated transcutaneous electrical stimulation. Furthermore, the median reflex amplitude for the 10 horses showed a tendency to decline over the stimulation train so temporal summation of afferent trigeminal inputs could not be observed. Therefore, the modulation of trigeminal nociceptive processing attributable to repeated Aδ fibre stimulations seems to differ from spinal processing of similar inputs as it seems to have an inhibitory rather than facilitatory effect. Further evaluation is necessary to highlight the underlying mechanism.

A Near-Real-Time Automatic Orbit Determination System for COSMIC and Its Follow-On Satellite Mission: Analysis of Orbit and Clock Errors on Radio Occultation

The COSMIC-2 mission is a follow-on mission of the Constellation Observing System for Meteorology, Ionosphere, and Climate (COSMIC) with an upgraded payload for improved radio occultation (RO) applications. The objective of this paper is to develop a near-real-time (NRT) orbit determination system, called NRT National Chiao Tung University (NCTU) system, to support COSMIC-2 in atmospheric applications and verify the orbit product of COSMIC. The system is capable of automatic determinations of the NRT GPS clocks and LEO orbit and clock. To assess the NRT (NCTU) system, we use eight days of COSMIC data (March 24-31, 2011), which contain a total of 331 GPS observation sessions and 12 393 RO observable files. The parallel scheduling for independent GPS and LEO estimations and automatic time matching improves the computational efficiency by 64% compared to the sequential scheduling. Orbit difference analyses suggest a 10-cm accuracy for the COSMIC orbits from the NRT (NCTU) system, and it is consistent as the NRT University Corporation for Atmospheric Research (URCA) system. The mean velocity accuracy from the NRT orbits of COSMIC is 0.168 mm/s, corresponding to an error of about 0.051 μrad in the bending angle. The rms differences in the NRT COSMIC clock and in GPS clocks between the NRT (NCTU) and the postprocessing products are 3.742 and 1.427 ns. The GPS clocks determined from a partial ground GPS network [from NRT (NCTU)] and a full one [from NRT (UCAR)] result in mean rms frequency stabilities of 6.1E-12 and 2.7E-12, respectively, corresponding to range fluctuations of 5.5 and 2.4 cm and bending angle errors of 3.75 and 1.66 μrad .

An optimized lead system for long-term esophageal electrocardiography

Long-term electrocardiography (ECG) featuring adequate atrial and ventricular signal quality is highly desirable. Routinely used surface leads are limited in atrial signal sensitivity and recording capability impeding complete ECG delineation, i.e. in the presence of supraventricular arrhythmias. Long-term esophageal ECG might overcome these limitations but requires a dedicated lead system and recorder design. To this end, we analysed multiple-lead esophageal ECGs with respect to signal quality by describing the ECG waves as a function of the insertion level, interelectrode distance, electrode shape and amplifier's input range. The results derived from clinical data show that two bipolar esophageal leads, an atrial lead with short (15 mm) interelectrode distance and a ventricular lead with long (80 mm) interelectrode distance provide non-inferior ventricular signal strength and superior atrial signal strength compared to standard surface lead II. High atrial signal slope in particular is observed with the atrial esophageal lead. The proposed esophageal lead system in combination with an increased recorder input range of ±20 mV minimizes signal loss due to excessive electrode motion typically observed in esophageal ECGs. The design proposal might help to standardize long-term esophageal ECG registrations and facilitate novel ECG classification systems based on the independent detection of ventricular and atrial electrical activity.

A Food Recognition System for Diabetic Patients based on an Optimized Bag of Features Model

Computer vision-based food recognition could be used to estimate a meal's carbohydrate content for diabetic patients. This study proposes a methodology for automatic food recognition, based on the Bag of Features (BoF) model. An extensive technical investigation was conducted for the identification and optimization of the best performing components involved in the BoF architecture, as well as the estimation of the corresponding parameters. For the design and evaluation of the prototype system, a visual dataset with nearly 5,000 food images was created and organized into 11 classes. The optimized system computes dense local features, using the scale-invariant feature transform on the HSV color space, builds a visual dictionary of 10,000 visual words by using the hierarchical k-means clustering and finally classifies the food images with a linear support vector machine classifier. The system achieved classification accuracy of the order of 78%, thus proving the feasibility of the proposed approach in a very challenging image dataset.