Showing papers by "Majid Mirmehdi published in 2015"

Bridging e-Health and the Internet of Things: The SPHERE Project

Abstract: There's a widely known need to revise current forms of healthcare provision. Of particular interest are sensing systems in the home, which have been central to several studies. This article presents an overview of this rapidly growing body of work, as well as the implications for machine learning, with an aim of uncovering the gap between the state of the art and the broad needs of healthcare services in ambient assisted living. Most approaches address specific healthcare concerns, which typically result in solutions that aren't able to support full-scale sensing and data analysis for a more generic healthcare service, but the approach in this article differs from seamlessly linking multimodel data-collecting infrastructure and data analytics together in an AAL platform. This article also outlines a multimodality sensor platform with heterogeneous network connectivity, which is under development in the sensor platform for healthcare in a residential environment (SPHERE) Interdisciplinary Research Collaboration (IRC).

Recognizing Text-Based Traffic Signs

Abstract: We propose a novel system for the automatic detection and recognition of text in traffic signs. Scene structure is used to define search regions within the image, in which traffic sign candidates are then found. Maximally stable extremal regions (MSERs) and hue, saturation, and value color thresholding are used to locate a large number of candidates, which are then reduced by applying constraints based on temporal and structural information. A recognition stage interprets the text contained within detected candidate regions. Individual text characters are detected as MSERs and are grouped into lines, before being interpreted using optical character recognition (OCR). Recognition accuracy is vastly improved through the temporal fusion of text results across consecutive frames. The method is comparatively evaluated and achieves an overall $F_{\rm measure}$ of 0.87.

Real-time RGB-D Tracking with Depth Scaling Kernelised Correlation Filters and Occlusion Handling

Abstract: We present a real-time RGB-D object tracker which manages occlusions and scale changes in a wide variety of scenarios. Its accuracy matches, and in many cases outperforms, state-of-the-art algorithms for precision and it far exceeds most in speed. We build our algorithm on the existing colour-only KCF tracker which uses the ‘kernel trick’ to extend correlation filters for fast tracking. We fuse colour and depth cues as the tracker’s features and exploit the depth data to both adjust a given target’s scale and to detect and manage occlusions in such a way as to maintain real-time performance, exceeding on average 35fps when benchmarked on two publicly available datasets. We make our easy-to-extend modularised code available to other researchers.

A multi-modal sensor infrastructure for healthcare in a residential environment

Abstract: Ambient Assisted Living (AAL) systems based on sensor technologies are seen as key enablers to an ageing society. However, most approaches in this space do not provide a truly generic ambient space - one that is not only capable of assisting people with diverse medical conditions, but can also recognise the habits of healthy habitants, as well as those with developing medical conditions. The recognition of Activities of Daily Living (ADL) is key to the understanding and provisioning of appropriate and efficient care. However, ADL recognition is particularly difficult to achieve in multi-resident spaces; especially with single-mode (albeit carefully crafted) solutions, which only have limited capabilities. To address these limitations we propose a multi-modal system architecture for AAL remote healthcare monitoring in the home, gathering information from multiple, diverse (sensor) data sources. In this paper we report on developments made to-date in various technical areas with respect to critical issues such as cost, power consumption, scalability, interoperability and privacy.

Skeleton-Free Body Pose Estimation from Depth Images for Movement Analysis

Abstract: In movement analysis frameworks, body pose may often be adequately represented in a simple, low-dimensional, and high-level space, while full body joints' locations constitute excessively redundant and complex information. We propose a method for estimating body pose in such high-level pose spaces, directly from a depth image and without relying on intermediate skeleton-based steps. Our method is based on a convolutional neural network (CNN) that maps the depth-silhouette of a person to its position in the pose space. We apply our method to a pose representation proposed in [16] that was initially built from skeleton data. We find our estimation of pose to be consistent with the original one, and to be suitable for use in the movement quality assessment framework of [16]. This opens the perspective of a wider applicability of the movement analysis method to movement types and view-angles that are not supported by its skeleton tracking algorithm.

A comparative home activity monitoring study using visual and inertial sensors

Abstract: Monitoring actions at home can provide essential information for rehabilitation management. This paper presents a comparative study and a dataset for the fully automated, sample-accurate recognition of common home actions in the living room environment using commercial-grade, inexpensive inertial and visual sensors. We investigate the practical home-use of body-worn mobile phone inertial sensors together with an Asus Xmotion RGB-Depth camera to achieve monitoring of daily living scenarios. To test this setup against realistic data, we introduce the challenging SPHERE-H130 action dataset containing 130 sequences of 13 household actions recorded in a home environment. We report automatic recognition results at maximal temporal resolution, which indicate that a vision-based approach outperforms accelerometer provided by two phone-based inertial sensors by an average of 14.85% accuracy for home actions. Further, we report improved accuracy of a vision-based approach over accelerometry on particularly challenging actions as well as when generalising across subjects.

Detection and Recognition of Painted Road Surface Markings

Abstract: A method for the automatic detection and recognition of text and symbols painted on the road surface is presented. Candidate regions are detected as maximally stable extremal regions (MSER) in a frame which has been transformed into an inverse perspective mapping (IPM) image, showing the road surface with the effects of perspective distortion removed. Detected candidates are then sorted into words and symbols, before they are interpreted using separate recognition stages. Symbol-based road markings are recognised using histogram of oriented gradient (HOG) features and support vector machines (SVM). Text-based road signs are recognised using a third-party optical character recognition (OCR) package, after application of a perspective correction stage. Matching of regions between frames, and temporal fusion of results is used to improve performance. The proposed method is validated using a data-set of videos, and achieves F-measures of 0.85 for text characters and 0.91 for symbols.

Remote pulmonary function testing using a depth sensor

Abstract: We propose a remote non-invasive approach to Pulmonary Function Testing using a time-of-flight depth sensor (Microsoft Kinect V2), and correlate our results to clinical-standard spirometry results. Given point clouds, we approximate 3D models of the subject's chest, estimate the volume throughout a sequence and construct volume-time and flow-time curves for two prevalent spirometry tests: Forced Vital Capacity and Slow Vital Capacity. From these, we compute clinical measures, such as FVC, FEV1, VC and IC. We correlate automatically extracted measures with clinical spirometry tests on 40 patients in an outpatient hospital setting. These demonstrate high within-test correlations.

Automatic Detection and Recognition of Symbols and Text on the Road Surface

Abstract: This paper presents a method for the automatic detection and recognition of text and symbols on the road surface, in the form of painted road markings. Candidates for road markings are detected as maximally stable extremal regions (MSER) in an inverse perspective mapping (IPM) transformed version of the image, which shows the road surface with the effects of perspective distortion removed. Separate recognition stages are then used to recognise words and symbols. Recognition of text-based regions is performed using a third-party optical character recognition (OCR) package, after application of a perspective correction stage. Symbol-based road markings are recognised by extracting histogram of oriented gradient (HOG) features, which are then classified using a support vector machine (SVM) classifier. The proposed method is validated using a data-set of videos, and achieves F-measures of 0.85 for text characters and 0.91 for symbols.

Cost-based Feature Transfer for Vehicle Occupant Classification

Abstract: Knowledge of human presence and interaction in a vehicle is of growing interest to vehicle manufacturers for design and safety purposes. We present a framework to perform the tasks of occupant detection and occupant classification for automatic child locks and airbag suppression. It operates for all passenger seats, using a single overhead camera. A transfer learning technique is introduced to make full use of training data from all seats whilst still maintaining some control over the bias, necessary for a system designed to penalize certain misclassifications more than others. An evaluation is performed on a challenging dataset with both weighted and unweighted classifiers, demonstrating the effectiveness of the transfer process.

Correspondence, Matching and Recognition

Abstract: The association of corresponding content across different visual representations and models is a fundamental task in many areas of computer vision. This special issue contains seven timely papers, all of which are concerned with solving a correspondence challenge, an associated matching or recognition task. The presented topics showcase some of the diversity in current computer vision research; they range widely from text recognition, motion segmentation, and cross-modal matching techniques to invariant descriptor construction, and aesthetic image analysis. Pons-Moll et al. present work, which aims at inferring dense data-to-model correspondences. In their paper “Metric Regression Forests for Correspondence Estimation” (doi:10. 1007/s11263-015-0818-9), they introduce a new decision forest training objective named Metric Space Information Gain (MSIG). They show that their methodology is a principled generalization of the proxy classification objective, which does not require an extrinsic isometric embedding of the model surface in Euclidean space. Backed by extensive experiments, the authors demonstrate that this leads to highly accurate associations, using few training images. Matching structures in cases where no one-to-one correspondences, but only relative pairing information is available is addressed in the paper “Relatively-Paired Space Analysis:

P83 Remote pulmonary function testing – computer gaming in the respiratory world

Abstract: Introduction Lung function testing by spirometry has remained unchanged for over 50 years, despite limitations including patient technique, discomfort, cost and training. Non-invasive, remote lung volume measurement is an alternative approach. This has been explored in the past using structured light, accelerometers and impedance pneumography, but these have often been costly and of limited validity. We use a novel approach to remote assessment (~2 metres) using a 3D time-of-flight depth camera – similar to those found in many home gaming consoles. This pilot developmental data was generated from patients in a clinical setting. Methods Patients were recruited from a general respiratory physiology laboratory. Spirometry was performed according to ATS/ERS standards using an unmodified pneumotachograph (nSpire Health, Longmont, CO, USA). A Kinect V2 time-of-flight depth sensor (Microsoft, Redmond, WA, USA) was used to reconstruct 3D models of the subject’s thorax to estimate volume-time and flow-time curves for both Forced and Slow Vital Capacity and their associated measurements (Figure 1, technical details in 1 ). These results were correlated with simultaneous recordings from the pneumotachograph, and error values calculated to assess the accuracy of the technique. Results Data were available from 53 patients, with 40 having usable data. Mean age 62.8 yrs (SD 16.2), BMI of 26.8 (SD 5.5). 41.5% male. 54.7% of patients had obstructive lung diseases, and 28.4% fibrotic lung disease. Mean FVC was 91.3% predicted (SD 26.4%), Mean FEV1 83.1% (SD 28.9%). The model estimates were highly correlated with spirometric values for FVC (λ = 0.999), FEV1 (λ = 0.947), VC (λ = 0.999), IC (λ = 0.997) and TV (λ = 0.962). Univariate analysis demonstrated no patient characteristics predictive of discrepancy from spirometric values for FVC or VC. Conclusions We describe a pilot data from the initial development of a new technique for non-invasively assessing lung volume and pulmonary function measurements. It correlates to within 30 ml for FVC and 10 ml for VC. This has a wide range of potential applications, including screening, home monitoring of respiratory disease, assessment of lung function in those unable to complete pneumotachygraphy and gating controls for radiological imaging techniques. Reference 1 Soleimani V, Mirmehdi M, Damen D, et al . Remote pulmonary function testing using a depth sensor. Biocas 2015