Search

Photoacoustic imaging is a promising technology for in vivo imaging. However, its imaging performance can be hampered by motion artifacts, especially when dealing with high-rate motion. In this paper, we propose an orthogonal motion correction method that utilizes cross-correlation along orthogonal scan directions to extract accurate motion displacements from the photoacoustic data. The extracted displacements are then applied to remove artifacts and compensate for motion-induced distortions. Phantom experiments demonstrate that the proposed method can extract the motion information and the structural similarity index measurement after correction is increased by 26.5% and 11.2% compared to no correction and the previous correction method. Then the effectiveness of our method is evaluated in vivo imaging of a mouse brain. Our method shows a stable and effective performance under high-rate motion. The high accuracy of the motion correction method makes it valuable in improving the accuracy of photoacoustic imaging. [ABSTRACT FROM AUTHOR]

We present a multi-stage neural network approach for 3D reconstruction of coronary artery trees from uncalibrated 2D X-ray angiography images. This method uses several binarized images from different angles to reconstruct a 3D coronary tree without any knowledge of image acquisition parameters. The method consists of a single backbone network and separate stages for vessel centerline and radius reconstruction. The output is an analytical matrix representation of the coronary tree suitable for downstream applications such as hemodynamic modeling of local vessel narrowing (i.e., stenosis). The network was trained using a dataset of synthetic coronary trees from a vessel generator informed by both clinical image data and literature values on coronary anatomy. Our multi-stage network achieved sub-pixel accuracy in reconstructing vessel radius (RMSE = 0.16 ± 0.07 mm) and stenosis radius (MAE = 0.27 ± 0.18 mm), the most important feature used to inform diagnostic decisions. The network also led to 52% and 38% reduction in vessel centerline reconstruction errors compared to a single-stage network and projective geometry-based methods, respectively. Our method demonstrated robustness to overcome challenges such as vessel foreshortening or overlap in the input images. This work is an important step towards automated analysis of anatomic and functional disease severity in the coronary arteries. [ABSTRACT FROM AUTHOR]

Human behaviour, along with any natural/biological behaviour, has varying degrees of intrinsic 'noise' or variability. Many studies have shown that the structure or patterning of this variability is sensitive to changes in task and constraint. Furthermore, two or more humans interacting together often begin to exhibit similar structures of behavioural variability (i.e., the patterning of their behavioural fluctuations becomes aligned or matched) independent of any moment-to-moment synchronization (termed complexity matching). However, much of the previous work has focused on a subset of simple or contrived behaviours within the context of highly controlled laboratory tasks. In the current study, individuals and pairs performed five self-paced (unsupervised), semi-structured activities around a university campus. Empatica E4 wristbands and iPhones were used to record the participants' behavioural activity via accelerometers and GPS. The results revealed that the structure of variability in naturalistic human behaviour co-varies with the task-goal constraints, and that the patterning of the behavioural fluctuations exhibited by co-acting individuals does become aligned during the performance of everyday activities. The results also revealed that the degree of complexity matching that occurred between pairs remained invariant across activity type, indicating that this measure could be employed as a robust, task-independent index of interpersonal behaviour.

Macrophages are heterogeneous innate immune cells that are functionally shaped by their surrounding microenvironment. Diverse macrophage populations have multifaceted differences related to their morphology, metabolism, expressed markers, and functions, where the identification of the different phenotypes is of an utmost importance in modelling immune response. While expressed markers are the most used signature to classify phenotypes, multiple reports indicate that macrophage morphology and autofluorescence are also valuable clues that can be used in the identification process. In this work, we investigated macrophage autofluorescence as a distinct feature for classifying six different macrophage phenotypes, namely: M0, M1, M2a, M2b, M2c, and M2d. The identification was based on extracted signals from multi-channel/multi-wavelength flow cytometer. To achieve the identification, we constructed a dataset containing 152,438 cell events each having a response vector of 45 optical signals fingerprint. Based on this dataset, we applied different supervised machine learning methods to detect phenotype specific fingerprint from the response vector, where the fully connected neural network architecture provided the highest classification accuracy of 75.8% for the six phenotypes compared simultaneously. Furthermore, by restricting the number of phenotypes in the experiment, the proposed framework produces higher classification accuracies, averaging 92.0%, 91.9%, 84.2%, and 80.4% for a pool of two, three, four, five phenotypes, respectively. These results indicate the potential of the intrinsic autofluorescence for classifying macrophage phenotypes, with the proposed method being quick, simple, and cost-effective way to accelerate the discovery of macrophage phenotypical diversity. [ABSTRACT FROM AUTHOR]

Stereotactic radiosurgery planning for cerebral arteriovenous malformations (AVM) is complicated by the variability in appearance of an AVM nidus across different imaging modalities. We developed a deep learning approach to automatically segment cerebrovascular-anatomical maps from multiple high-resolution magnetic resonance imaging/angiography (MRI/MRA) sequences in AVM patients, with the goal of facilitating target delineation. Twenty-three AVM patients who were evaluated for radiosurgery and underwent multi-parametric MRI/MRA were included. A hybrid semi-automated and manual approach was used to label MRI/MRAs with arteries, veins, brain parenchyma, cerebral spinal fluid (CSF), and embolized vessels. Next, these labels were used to train a convolutional neural network to perform this task. Imaging from 17 patients (6362 image slices) was used for training, and 6 patients (1224 slices) for validation. Performance was evaluated by Dice Similarity Coefficient (DSC). Classification performance was good for arteries, veins, brain parenchyma, and CSF, with DSCs of 0.86, 0.91, 0.98, and 0.91, respectively in the validation image set. Performance was lower for embolized vessels, with a DSC of 0.75. This demonstrates the proof of principle that accurate, high-resolution cerebrovascular-anatomical maps can be generated from multiparametric MRI/MRA. Clinical validation of their utility in radiosurgery planning is warranted. [ABSTRACT FROM AUTHOR]

Malnutrition is a multidomain problem affecting 54% of older adults in long-term care (LTC). Monitoring nutritional intake in LTC is laborious and subjective, limiting clinical inference capabilities. Recent advances in automatic image-based food estimation have not yet been evaluated in LTC settings. Here, we describe a fully automatic imaging system for quantifying food intake. We propose a novel deep convolutional encoder-decoder food network with depth-refinement (EDFN-D) using an RGB-D camera for quantifying a plate's remaining food volume relative to reference portions in whole and modified texture foods. We trained and validated the network on the pre-labelled UNIMIB2016 food dataset and tested on our two novel LTC-inspired plate datasets (689 plate images, 36 unique foods). EDFN-D performed comparably to depth-refined graph cut on IOU (0.879 vs. 0.887), with intake errors well below typical 50% (mean percent intake error: - 4.2 %). We identify how standard segmentation metrics are insufficient due to visual-volume discordance, and include volume disparity analysis to facilitate system trust. This system provides improved transparency, approximates human assessors with enhanced objectivity, accuracy, and precision while avoiding hefty semi-automatic method time requirements. This may help address short-comings currently limiting utility of automated early malnutrition detection in resource-constrained LTC and hospital settings. [ABSTRACT FROM AUTHOR]

Psoriasis is a chronic inflammatory skin disease that occurs in various forms throughout the body and is associated with certain conditions such as heart disease, diabetes, and depression. The psoriasis area severity index (PASI) score, a tool used to evaluate the severity of psoriasis, is currently used in clinical trials and clinical research. The determination of severity is based on the subjective judgment of the clinician. Thus, the disease evaluation deviations are induced. Therefore, we propose optimal algorithms that can effectively segment the lesion area and classify the severity. In addition, a new dataset on psoriasis was built, including patch images of erythema and scaling. We performed psoriasis lesion segmentation and classified the disease severity. In addition, we evaluated the best-performing segmentation method and classifier and analyzed features that are highly related to the severity of psoriasis. In conclusion, we presented the optimal techniques for evaluating the severity of psoriasis. Our newly constructed dataset improved the generalization performance of psoriasis diagnosis and evaluation. It proposed an optimal system for specific evaluation indicators of the disease and a quantitative PASI scoring method. The proposed system can help to evaluate the severity of localized psoriasis more accurately. [ABSTRACT FROM AUTHOR]

Coronary Artery Disease (CAD) is commonly diagnosed using X-ray angiography, in which images are taken as radio-opaque dye is flushed through the coronary vessels to visualize the severity of vessel narrowing, or stenosis. Cardiologists typically use visual estimation to approximate the percent diameter reduction of the stenosis, and this directs therapies like stent placement. A fully automatic method to segment the vessels would eliminate potential subjectivity and provide a quantitative and systematic measurement of diameter reduction. Here, we have designed a convolutional neural network, AngioNet, for vessel segmentation in X-ray angiography images. The main innovation in this network is the introduction of an Angiographic Processing Network (APN) which significantly improves segmentation performance on multiple network backbones, with the best performance using Deeplabv3+ (Dice score 0.864, pixel accuracy 0.983, sensitivity 0.918, specificity 0.987). The purpose of the APN is to create an end-to-end pipeline for image pre-processing and segmentation, learning the best possible pre-processing filters to improve segmentation. We have also demonstrated the interchangeability of our network in measuring vessel diameter with Quantitative Coronary Angiography. Our results indicate that AngioNet is a powerful tool for automatic angiographic vessel segmentation that could facilitate systematic anatomical assessment of coronary stenosis in the clinical workflow. [ABSTRACT FROM AUTHOR]

Properties of solid-state materials depend on their crystal structures. In solid solution high entropy alloy (HEA), its mechanical properties such as strength and ductility depend on its phase. Therefore, the crystal structure prediction should be preceded to find new functional materials. Recently, the machine learning-based approach has been successfully applied to the prediction of structural phases. However, since about 80% of the data set is used as a training set in machine learning, it is well known that it requires vast cost for preparing a dataset of multi-element alloy as training. In this work, we develop an efficient approach to predicting the multi-element alloys' structural phases without preparing a large scale of the training dataset. We demonstrate that our method trained from binary alloy dataset can be applied to the multi-element alloys' crystal structure prediction by designing a transformation module from raw features to expandable form. Surprisingly, without involving the multi-element alloys in the training process, we obtain an accuracy, 80.56% for the phase of the multi-element alloy and 84.20% accuracy for the phase of HEA. It is comparable with the previous machine learning results. Besides, our approach saves at least three orders of magnitude computational cost for HEA by employing expandable features. We suggest that this accelerated approach can be applied to predicting various structural properties of multi-elements alloys that do not exist in the current structural database. [ABSTRACT FROM AUTHOR]

Accurately segmenting foods from optical images is a challenging task, yet becoming possible with the help of recent advances in Deep Learning based solutions. Automated identification of food items opens up possibilities of useful applications like nutrition intake monitoring. Given large variations in food choices, Deep Learning based solutions still struggle to generate human level accuracy. In this work, we propose a novel Sequential Transfer Learning method using Hierarchical Clustering. This novel approach simulates a step by step problem solving framework based on clustering of similar types of foods. The proposed approach provides up to 6% gain in accuracy compared to traditional network training and generated a robust model performing better in challenging unseen cases. This approach is also tested for segmenting foods in Danish school children meals for dietary intake monitoring as an application. [ABSTRACT FROM AUTHOR]

Capsule endoscopy (CE) is a widely used, minimally invasive alternative to traditional endoscopy that allows visualisation of the entire small intestine. Patient preparation can help to obtain a cleaner intestine and thus better visibility in the resulting videos. However, studies on the most effective preparation method are conflicting due to the absence of objective, automatic cleanliness evaluation methods. In this work, we aim to provide such a method capable of presenting results on an intuitive scale, with a relatively light-weight novel convolutional neural network architecture at its core. We trained our model using 5-fold cross-validation on an extensive data set of over 50,000 image patches, collected from 35 different CE procedures, and compared it with state-of-the-art classification methods. From the patch classification results, we developed a method to automatically estimate pixel-level probabilities and deduce cleanliness evaluation scores through automatically learnt thresholds. We then validated our method in a clinical setting on 30 newly collected CE videos, comparing the resulting scores to those independently assigned by human specialists. We obtained the highest classification accuracy for the proposed method (95.23%), with significantly lower average prediction times than for the second-best method. In the validation of our method, we found acceptable agreement with two human specialists compared to interhuman agreement, showing its validity as an objective evaluation method. [ABSTRACT FROM AUTHOR]

The increasingly ubiquitous use of embedded devices calls for autonomous optimizations of sensor performance with meager computing resources. Due to the heavy computing needs, such optimization is rarely performed, and almost never carried out on-the-fly, resulting in a vast underutilization of deployed assets. Aiming at improving the measurement efficiency, we show an OED (Optimal Experimental Design) routine where quantities of interest of probable samples are partitioned into distinctive classes, with the corresponding sensor signals learned by supervised learning models. The trained models, digesting the compressed live data, are subsequently executed at the constrained device for continuous classification and optimization of measurements. We demonstrate the closed-loop method with multidimensional NMR (Nuclear Magnetic Resonance) relaxometry, an analytical technique seeing a substantial growth of field applications in recent years, on a wide range of complex fluids. The realtime portion of the procedure demands minimal computing load, and is ideally suited for instruments that are widely used in remote sensing and IoT networks. [ABSTRACT FROM AUTHOR]

Imaging software have become critical tools in the diagnosis and the treatment of abdominal aortic aneurysms (AAA). The aim of this study was to develop a fully automated software system to enable a fast and robust detection of the vascular system and the AAA. The software was designed from a dataset of injected CT-scans images obtained from 40 patients with AAA. Pre-processing steps were performed to reduce the noise of the images using image filters. The border propagation based method was used to localize the aortic lumen. An online error detection was implemented to correct errors due to the propagation in anatomic structures with similar pixel value located close to the aorta. A morphological snake was used to segment 2D or 3D regions. The software allowed an automatic detection of the aortic lumen and the AAA characteristics including the presence of thrombus and calcifications. 2D and 3D reconstructions visualization were available to ease evaluation of both algorithm precision and AAA properties. By enabling a fast and automated detailed analysis of the anatomic characteristics of the AAA, this software could be useful in clinical practice and research and be applied in a large dataset of patients. [ABSTRACT FROM AUTHOR]

Myocardial tracking and strain estimation can non-invasively assess cardiac functioning using subject-specific MRI. As the left-ventricle does not have a uniform shape and functioning from base to apex, the development of 3D MRI has provided opportunities for simultaneous 3D tracking, and 3D strain estimation. We have extended a Local Weighted Mean (LWM) transformation function for 3D, and incorporated in a Hierarchical Template Matching model to solve 3D myocardial tracking and strain estimation problem. The LWM does not need to solve a large system of equations, provides smooth displacement of myocardial points, and adapt local geometric differences in images. Hence, 3D myocardial tracking can be performed with 1.49 mm median error, and without large error outliers. The maximum error of tracking is up to 24% reduced compared to benchmark methods. Moreover, the estimated strain can be insightful to improve 3D imaging protocols, and the computer code of LWM could also be useful for geo-spatial and manufacturing image analysis researchers. [ABSTRACT FROM AUTHOR]