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.

Non-invasive and non-occlusive blood pressure estimation via a chest sensor

The clinical demand for a device to monitor Blood Pressure (BP) in ambulatory scenarios with minimal use of inflation cuffs is increasing. Based on the so-called Pulse Wave Velocity (PWV) principle, this paper introduces and evaluates a novel concept of BP monitor that can be fully integrated within a chest sensor. After a preliminary calibration, the sensor provides non-occlusive beat-by-beat estimations of Mean Arterial Pressure (MAP) by measuring the Pulse Transit Time (PTT) of arterial pressure pulses travelling from the ascending aorta towards the subcutaneous vasculature of the chest. In a cohort of 15 healthy male subjects, a total of 462 simultaneous readings consisting of reference MAP and chest PTT were acquired. Each subject was recorded at three different days: D, D+3 and D+14. Overall, the implemented protocol induced MAP values to range from 80 ± 6 mmHg in baseline, to 107 ± 9 mmHg during isometric handgrip maneuvers. Agreement between reference and chest-sensor MAP values was tested by using intraclass correlation coefficient (ICC = 0.78) and Bland-Altman analysis (mean error = 0.7 mmHg, standard deviation = 5.1 mmHg). The cumulative percentage of MAP values provided by the chest sensor falling within a range of ±5 mmHg compared to reference MAP readings was of 70%, within ±10 mmHg was of 91%, and within ±15mmHg was of 98%. These results point at the fact that the chest sensor complies with the British Hypertension Society (BHS) requirements of Grade A BP monitors, when applied to MAP readings. Grade A performance was maintained even two weeks after having performed the initial subject-dependent calibration. In conclusion, this paper introduces a sensor and a calibration strategy to perform MAP measurements at the chest. The encouraging performance of the presented technique paves the way towards an ambulatory-compliant, continuous and non-occlusive BP monitoring system.

The human vestibular cortex revealed by coordinate-based activation likelihood estimation meta-analysis.

The vestibular system contributes to the control of posture and eye movements and is also involved in various cognitive functions including spatial navigation and memory. These functions are subtended by projections to a vestibular cortex, whose exact location in the human brain is still a matter of debate (Lopez and Blanke, 2011). The vestibular cortex can be defined as the network of all cortical areas receiving inputs from the vestibular system, including areas where vestibular signals influence the processing of other sensory (e.g. somatosensory and visual) and motor signals. Previous neuroimaging studies used caloric vestibular stimulation (CVS), galvanic vestibular stimulation (GVS), and auditory stimulation (clicks and short-tone bursts) to activate the vestibular receptors and localize the vestibular cortex. However, these three methods differ regarding the receptors stimulated (otoliths, semicircular canals) and the concurrent activation of the tactile, thermal, nociceptive and auditory systems. To evaluate the convergence between these methods and provide a statistical analysis of the localization of the human vestibular cortex, we performed an activation likelihood estimation (ALE) meta-analysis of neuroimaging studies using CVS, GVS, and auditory stimuli. We analyzed a total of 352 activation foci reported in 16 studies carried out in a total of 192 healthy participants. The results reveal that the main regions activated by CVS, GVS, or auditory stimuli were located in the Sylvian fissure, insula, retroinsular cortex, fronto-parietal operculum, superior temporal gyrus, and cingulate cortex. Conjunction analysis indicated that regions showing convergence between two stimulation methods were located in the median (short gyrus III) and posterior (long gyrus IV) insula, parietal operculum and retroinsular cortex (Ri). The only area of convergence between all three methods of stimulation was located in Ri. The data indicate that Ri, parietal operculum and posterior insula are vestibular regions where afferents converge from otoliths and semicircular canals, and may thus be involved in the processing of signals informing about body rotations, translations and tilts. Results from the meta-analysis are in agreement with electrophysiological recordings in monkeys showing main vestibular projections in the transitional zone between Ri, the insular granular field (Ig), and SII.

A web-based non-intrusive ambient system to measure and classify activities of daily living.

BACKGROUND The number of older adults in the global population is increasing. This demographic shift leads to an increasing prevalence of age-associated disorders, such as Alzheimer's disease and other types of dementia. With the progression of the disease, the risk for institutional care increases, which contrasts with the desire of most patients to stay in their home environment. Despite doctors' and caregivers' awareness of the patient's cognitive status, they are often uncertain about its consequences on activities of daily living (ADL). To provide effective care, they need to know how patients cope with ADL, in particular, the estimation of risks associated with the cognitive decline. The occurrence, performance, and duration of different ADL are important indicators of functional ability. The patient's ability to cope with these activities is traditionally assessed with questionnaires, which has disadvantages (eg, lack of reliability and sensitivity). Several groups have proposed sensor-based systems to recognize and quantify these activities in the patient's home. Combined with Web technology, these systems can inform caregivers about their patients in real-time (e.g., via smartphone). OBJECTIVE We hypothesize that a non-intrusive system, which does not use body-mounted sensors, video-based imaging, and microphone recordings would be better suited for use in dementia patients. Since it does not require patient's attention and compliance, such a system might be well accepted by patients. We present a passive, Web-based, non-intrusive, assistive technology system that recognizes and classifies ADL. METHODS The components of this novel assistive technology system were wireless sensors distributed in every room of the participant's home and a central computer unit (CCU). The environmental data were acquired for 20 days (per participant) and then stored and processed on the CCU. In consultation with medical experts, eight ADL were classified. RESULTS In this study, 10 healthy participants (6 women, 4 men; mean age 48.8 years; SD 20.0 years; age range 28-79 years) were included. For explorative purposes, one female Alzheimer patient (Montreal Cognitive Assessment score=23, Timed Up and Go=19.8 seconds, Trail Making Test A=84.3 seconds, Trail Making Test B=146 seconds) was measured in parallel with the healthy subjects. In total, 1317 ADL were performed by the participants, 1211 ADL were classified correctly, and 106 ADL were missed. This led to an overall sensitivity of 91.27% and a specificity of 92.52%. Each subject performed an average of 134.8 ADL (SD 75). CONCLUSIONS The non-intrusive wireless sensor system can acquire environmental data essential for the classification of activities of daily living. By analyzing retrieved data, it is possible to distinguish and assign data patterns to subjects' specific activities and to identify eight different activities in daily living. The Web-based technology allows the system to improve care and provides valuable information about the patient in real-time.

Long-distance photogrammetric trait estimation in free-ranging animals: A new approach

The acquisition of accurate information on the size of traits in animals is fundamental for the study of animal ecology and evolution and their management. We demonstrate how morphological traits of free-ranging animals can reliably be estimated on very large observation distances of several hundred meters by the use of ordinary digital photographic equipment and simple photogrammetric software. In our study, we estimated the length of horn annuli in free-ranging male Alpine ibex (Capra ibex) by taking already measured horn annuli of conspecifics on the same photographs as scaling units. Comparisons with hand-measured horn annuli lengths and repeatability analyses revealed a high accuracy of the photogrammetric estimates. If length estimations of specific horn annuli are based on multiple photographs measurement errors of <5.5 mm can be expected. In the current study the application of the described photogrammetric procedure increased the sample size of animals with known horn annuli length by an additional 104%. The presented photogrammetric procedure is of broad applicability and represents an easy, robust and cost-efficient method for the measuring of individuals in populations where animals are hard to capture or to approach.

DYPTOP: a cost-efficient TOPMODEL implementation to simulate sub-grid spatio-temporal dynamics of global wetlands and peatlands

Simulating the spatio-temporal dynamics of inundation is key to understanding the role of wetlands under past and future climate change. Earlier modelling studies have mostly relied on fixed prescribed peatland maps and inundation time series of limited temporal coverage. Here, we describe and assess the the Dynamical Peatland Model Based on TOPMODEL (DYPTOP), which predicts the extent of inundation based on a computationally efficient TOPMODEL implementation. This approach rests on an empirical, grid-cell-specific relationship between the mean soil water balance and the flooded area. DYPTOP combines the simulated inundation extent and its temporal persistency with criteria for the ecosystem water balance and the modelled peatland-specific soil carbon balance to predict the global distribution of peatlands. We apply DYPTOP in combination with the LPX-Bern DGVM and benchmark the global-scale distribution, extent, and seasonality of inundation against satellite data. DYPTOP successfully predicts the spatial distribution and extent of wetlands and major boreal and tropical peatland complexes and reveals the governing limitations to peatland occurrence across the globe. Peatlands covering large boreal lowlands are reproduced only when accounting for a positive feedback induced by the enhanced mean soil water holding capacity in peatland-dominated regions. DYPTOP is designed to minimize input data requirements, optimizes computational efficiency and allows for a modular adoption in Earth system models.

Comparison of (31)P saturation and inversion magnetization transfer in human liver and skeletal muscle using a clinical MR system and surface coils.

(31)P MRS magnetization transfer ((31)P-MT) experiments allow the estimation of exchange rates of biochemical reactions, such as the creatine kinase equilibrium and adenosine triphosphate (ATP) synthesis. Although various (31)P-MT methods have been successfully used on isolated organs or animals, their application on humans in clinical scanners poses specific challenges. This study compared two major (31)P-MT methods on a clinical MR system using heteronuclear surface coils. Although saturation transfer (ST) is the most commonly used (31)P-MT method, sequences such as inversion transfer (IT) with short pulses might be better suited for the specific hardware and software limitations of a clinical scanner. In addition, small NMR-undetectable metabolite pools can transfer MT to NMR-visible pools during long saturation pulses, which is prevented with short pulses. (31)P-MT sequences were adapted for limited pulse length, for heteronuclear transmit-receive surface coils with inhomogeneous B1 , for the need for volume selection and for the inherently low signal-to-noise ratio (SNR) on a clinical 3-T MR system. The ST and IT sequences were applied to skeletal muscle and liver in 10 healthy volunteers. Monte-Carlo simulations were used to evaluate the behavior of the IT measurements with increasing imperfections. In skeletal muscle of the thigh, ATP synthesis resulted in forward reaction constants (k) of 0.074 ± 0.022 s(-1) (ST) and 0.137 ± 0.042 s(-1) (IT), whereas the creatine kinase reaction yielded 0.459 ± 0.089 s(-1) (IT). In the liver, ATP synthesis resulted in k = 0.267 ± 0.106 s(-1) (ST), whereas the IT experiment yielded no consistent results. ST results were close to literature values; however, the IT results were either much larger than the corresponding ST values and/or were widely scattered. To summarize, ST and IT experiments can both be implemented on a clinical body scanner with heteronuclear transmit-receive surface coils; however, ST results are much more robust against experimental imperfections than the current implementation of IT.

Automatic estimation of noise parameters in Fourier-domain optical coherence tomography cross sectional images using statistical information

We present an application and sample independent method for the automatic discrimination of noise and signal in optical coherence tomography Bscans. The proposed algorithm models the observed noise probabilistically and allows for a dynamic determination of image noise parameters and the choice of appropriate image rendering parameters. This overcomes the observer variability and the need for a priori information about the content of sample images, both of which are challenging to estimate systematically with current systems. As such, our approach has the advantage of automatically determining crucial parameters for evaluating rendered image quality in a systematic and task independent way. We tested our algorithm on data from four different biological and nonbiological samples (index finger, lemon slices, sticky tape, and detector cards) acquired with three different experimental spectral domain optical coherence tomography (OCT) measurement systems including a swept source OCT. The results are compared to parameters determined manually by four experienced OCT users. Overall, our algorithm works reliably regardless of which system and sample are used and estimates noise parameters in all cases within the confidence interval of those found by observers.

Computer Vision-Based Carbohydrate Estimation for Type 1 Diabetic Patients Using Smartphones

Background: Individuals with type 1 diabetes (T1D) have to count the carbohydrates (CHOs) of their meal to estimate the prandial insulin dose needed to compensate for the meal’s effect on blood glucose levels. CHO counting is very challenging but also crucial, since an error of 20 grams can substantially impair postprandial control. Method: The GoCARB system is a smartphone application designed to support T1D patients with CHO counting of nonpacked foods. In a typical scenario, the user places a reference card next to the dish and acquires 2 images with his/her smartphone. From these images, the plate is detected and the different food items on the plate are automatically segmented and recognized, while their 3D shape is reconstructed. Finally, the food volumes are calculated and the CHO content is estimated by combining the previous results and using the USDA nutritional database. Results: To evaluate the proposed system, a set of 24 multi-food dishes was used. For each dish, 3 pairs of images were taken and for each pair, the system was applied 4 times. The mean absolute percentage error in CHO estimation was 10 ± 12%, which led to a mean absolute error of 6 ± 8 CHO grams for normal-sized dishes. Conclusion: The laboratory experiments demonstrated the feasibility of the GoCARB prototype system since the error was below the initial goal of 20 grams. However, further improvements and evaluation are needed prior launching a system able to meet the inter- and intracultural eating habits.

Range-based Weighted-likelihood Particle Filter for RSS-based Indoor Tracking

Attractive business cases in various application fields contribute to the sustained long-term interest in indoor localization and tracking by the research community. Location tracking is generally treated as a dynamic state estimation problem, consisting of two steps: (i) location estimation through measurement, and (ii) location prediction. For the estimation step, one of the most efficient and low-cost solutions is Received Signal Strength (RSS)-based ranging. However, various challenges - unrealistic propagation model, non-line of sight (NLOS), and multipath propagation - are yet to be addressed. Particle filters are a popular choice for dealing with the inherent non-linearities in both location measurements and motion dynamics. While such filters have been successfully applied to accurate, time-based ranging measurements, dealing with the more error-prone RSS based ranging is still challenging. In this work, we address the above issues with a novel, weighted likelihood, bootstrap particle filter for tracking via RSS-based ranging. Our filter weights the individual likelihoods from different anchor nodes exponentially, according to the ranging estimation. We also employ an improved propagation model for more accurate RSS-based ranging, which we suggested in recent work. We implemented and tested our algorithm in a passive localization system with IEEE 802.15.4 signals, showing that our proposed solution largely outperforms a traditional bootstrap particle filter.

A cost-effective surgical navigation solution for periacetabular osteotomy (PAO) surgery

In this chapter a low-cost surgical navigation solution for periacetabular osteotomy (PAO) surgery is described. Two commercial inertial measurement units (IMU, Xsens Technologies, The Netherlands), are attached to a patient’s pelvis and to the acetabular fragment, respectively. Registration of the patient with a pre-operatively acquired computer model is done by recording the orientation of the patient’s anterior pelvic plane (APP) using one IMU. A custom-designed device is used to record the orientation of the APP in the reference coordinate system of the IMU. After registration, the two sensors are mounted to the patient’s pelvis and acetabular fragment, respectively. Once the initial position is recorded, the orientation is measured and displayed on a computer screen. A patient-specific computer model generated from a pre-operatively acquired computed tomography (CT) scan is used to visualize the updated orientation of the acetabular fragment. Experiments with plastic bones (7 hip joints) performed in an operating room comparing a previously developed optical navigation system with our inertial-based navigation system showed no statistical difference on the measurement of acetabular component reorientation (anteversion and inclination). In six out of seven hip joints the mean absolute difference was below five degrees for both anteversion and inclination.

A cost-effective surgical navigation solution for periacetabular osteotomy (PAO) surgery.

PURPOSE To evaluate a low-cost, inertial sensor-based surgical navigation solution for periacetabular osteotomy (PAO) surgery without the line-of-sight impediment. METHODS Two commercial inertial measurement units (IMU, Xsens Technologies, The Netherlands), are attached to a patient's pelvis and to the acetabular fragment, respectively. Registration of the patient with a pre-operatively acquired computer model is done by recording the orientation of the patient's anterior pelvic plane (APP) using one IMU. A custom-designed device is used to record the orientation of the APP in the reference coordinate system of the IMU. After registration, the two sensors are mounted to the patient's pelvis and acetabular fragment, respectively. Once the initial position is recorded, the orientation is measured and displayed on a computer screen. A patient-specific computer model generated from a pre-operatively acquired computed tomography scan is used to visualize the updated orientation of the acetabular fragment. RESULTS Experiments with plastic bones (eight hip joints) performed in an operating room comparing a previously developed optical navigation system with our inertial-based navigation system showed no statistically significant difference on the measurement of acetabular component reorientation. In all eight hip joints the mean absolute difference was below four degrees. CONCLUSION Using two commercially available inertial measurement units we show that it is possible to accurately measure the orientation (inclination and anteversion) of the acetabular fragment during PAO surgery and therefore to successfully eliminate the line-of-sight impediment that optical navigation systems have.

A clinically applicable laser-based image-guided system for laparoscopic liver procedures

PURPOSE Laser range scanners (LRS) allow performing a surface scan without physical contact with the organ, yielding higher registration accuracy for image-guided surgery (IGS) systems. However, the use of LRS-based registration in laparoscopic liver surgery is still limited because current solutions are composed of expensive and bulky equipment which can hardly be integrated in a surgical scenario. METHODS In this work, we present a novel LRS-based IGS system for laparoscopic liver procedures. A triangulation process is formulated to compute the 3D coordinates of laser points by using the existing IGS system tracking devices. This allows the use of a compact and cost-effective LRS and therefore facilitates the integration into the laparoscopic setup. The 3D laser points are then reconstructed into a surface to register to the preoperative liver model using a multi-level registration process. RESULTS Experimental results show that the proposed system provides submillimeter scanning precision and accuracy comparable to those reported in the literature. Further quantitative analysis shows that the proposed system is able to achieve a patient-to-image registration accuracy, described as target registration error, of [Formula: see text]. CONCLUSIONS We believe that the presented approach will lead to a faster integration of LRS-based registration techniques in the surgical environment. Further studies will focus on optimizing scanning time and on the respiratory motion compensation.