Multi-modal object tracking using dynamic performance metrics

Intelligent surveillance systems typically use a single visual spectrum modality for their input. These systems work well in controlled conditions, but often fail when lighting is poor, or environmental effects such as shadows, dust or smoke are present. Thermal spectrum imagery is not as susceptible to environmental effects, however thermal imaging sensors are more sensitive to noise and they are only gray scale, making distinguishing between objects difficult. Several approaches to combining the visual and thermal modalities have been proposed, however they are limited by assuming that both modalities are perfuming equally well. When one modality fails, existing approaches are unable to detect the drop in performance and disregard the under performing modality. In this paper, a novel middle fusion approach for combining visual and thermal spectrum images for object tracking is proposed. Motion and object detection is performed on each modality and the object detection results for each modality are fused base on the current performance of each modality. Modality performance is determined by comparing the number of objects tracked by the system with the number detected by each mode, with a small allowance made for objects entering and exiting the scene. The tracking performance of the proposed fusion scheme is compared with performance of the visual and thermal modes individually, and a baseline middle fusion scheme. Improvement in tracking performance using the proposed fusion approach is demonstrated. The proposed approach is also shown to be able to detect the failure of an individual modality and disregard its results, ensuring performance is not degraded in such situations.

An improved modal strain energy method for damage assessment

The modal strain energy method, which depends on the vibration characteristics of the structure, has been reasonably successful in identifying and localising damage in the structure. However, existing strain energy methods require the first few modes to be measured to provide meaningful damage detection. Use of individual modes with existing strain energy methods may indicate false alarms or may not detect the damage at or near the nodal points. This paper proposes a new modal strain energy based damage index which can detect and localize the damage using any one of the modes measured and illustrates its application for beam structures. It becomes evident that the proposed strain energy based damage index also has potential for damage quantification.

Response shift, recall bias and their effect on measuring change in health-related quality of life amongst older hospital patients

Background: Assessments of change in subjective patient reported outcomes such as health-related quality of life (HRQoL) are a key component of many clinical and research evaluations. However, conventional longitudinal evaluation of change may not agree with patient perceived change if patients' understanding of the subjective construct under evaluation changes over time (response shift) or if patients' have inaccurate recollection (recall bias). This study examined whether older adults' perception of change is in agreement with conventional longitudinal evaluation of change in their HRQoL over the duration of their hospital stay. It also investigated this level of agreement after adjusting patient perceived change for recall bias that patients may have experienced. Methods: A prospective longitudinal cohort design nested within a larger randomised controlled trial was implemented. 103 hospitalised older adults participated in this investigation at a tertiary hospital facility. The EQ-5D utility and Visual Analogue Scale (VAS) scores were used to evaluate HRQoL. Participants completed EQ-5D reports as soon as they were medically stable (within three days of admission) then again immediately prior to discharge. Three methods of change score calculation were used (conventional change, patient perceived change and patient perceived change adjusted for recall bias). Agreement was primarily investigated using intraclass correlation coefficients (ICC) and limits of agreement. Results: Overall 101 (98%) participants completed both admission and discharge assessments. The mean (SD) age was 73.3 (11.2). The median (IQR) length of stay was 38 (20-60) days. For agreement between conventional longitudinal change and patient perceived change: ICCs were 0.34 and 0.40 for EQ-5D utility and VAS respectively. For agreement between conventional longitudinal change and patient perceived change adjusted for recall bias: ICCs were 0.98 and 0.90 respectively. Discrepancy between conventional longitudinal change and patient perceived change was considered clinically meaningful for 84 (83.2%) of participants, after adjusting for recall bias this reduced to 8 (7.9%). Conclusions: Agreement between conventional change and patient perceived change was not strong. A large proportion of this disagreement could be attributed to recall bias. To overcome the invalidating effect of response shift (on conventional change) and recall bias (on patient perceived change) a method of adjusting patient perceived change for recall bias has been described.

Modal parameter estimation in the presence of non-linearities

This paper investigates the use of time-frequency techniques to assist in the estimation of power system modes which are resolvable by a Digital Fourier Transform (DFT). The limitations of linear estimation techniques in the presence of large disturbances which excite system non-linearities, particularly the swing equation non-linearity are shown. Where a nonlinearity manifests itself as time varying modal frequencies the Wigner-Ville Distribution (WVD) is used to describe the variation in modal frequencies and construct a window over which standard linear estimation techniques can be used. The error obtained even in the presence of multiple resolvable modes is better than 2%.

Quantifying axial deformations of core shear walls using modal parameters

Differential axial deformation between column elements and shear wall elements of cores increase with building height and geometric complexity. Adverse effects due to the differential axial deformation reduce building performance and life time serviceability. Quantifying axial deformations using ambient measurements from vibrating wire, external mechanical and electronic strain gauges in order to acquire adequate provisions to mitigate the adverse effects is well established method. However, these gauges require installing in or on elements to acquire continuous measurements and hence use of these gauges is uneconomical and inconvenient. This motivates to develop a method to quantify the axial deformations. This paper proposes an innovative method based on modal parameters to quantify axial deformations of shear wall elements in cores of buildings. Capabilities of the method are presented though an illustrative example.

Multi-modal summarization of key events and top players in sports tournament videos

To detect and annotate the key events of live sports videos, we need to tackle the semantic gaps of audio-visual information. Previous work has successfully extracted semantic from the time-stamped web match reports, which are synchronized with the video contents. However, web and social media articles with no time-stamps have not been fully leveraged, despite they are increasingly used to complement the coverage of major sporting tournaments. This paper aims to address this limitation using a novel multimodal summarization framework that is based on sentiment analysis and players' popularity. It uses audiovisual contents, web articles, blogs, and commentators' speech to automatically annotate and visualize the key events and key players in a sports tournament coverage. The experimental results demonstrate that the automatically generated video summaries are aligned with the events identified from the official website match reports.

The use of temporal speech and lip information for multi-modal speaker identification via multi-stream HMMs

Investigates the use of temporal lip information, in conjunction with speech information, for robust, text-dependent speaker identification. We propose that significant speaker-dependent information can be obtained from moving lips, enabling speaker recognition systems to be highly robust in the presence of noise. The fusion structure for the audio and visual information is based around the use of multi-stream hidden Markov models (MSHMM), with audio and visual features forming two independent data streams. Recent work with multi-modal MSHMMs has been performed successfully for the task of speech recognition. The use of temporal lip information for speaker identification has been performed previously (T.J. Wark et al., 1998), however this has been restricted to output fusion via single-stream HMMs. We present an extension to this previous work, and show that a MSHMM is a valid structure for multi-modal speaker identification

Characteristics of modal sound radiation of finite cylindrical shells

Characteristics of modal sound radiation of finite cylindrical shells are studied using finite element and boundary element methods in this paper. In the low frequency range, modal radiation efficiencies of finite cylindrical shells are found to asymptotically approach those of the corresponding infinite cylindrical shell when structural trace wavelengths of the cylindrical shells are greater than the acoustic wavelength. Modal radiation efficiencies for each group of modes having the same circumferential modal index decrease as the axial modal index increases. They converge to each other when the axial trace wavelength is much greater than the circumferential trace wavelength. The mechanism leading to lower radiation efficiency of modes with higher circumferential modal index of short cylinders is explained. Similar to those of flat plate panels, change in slope or waviness is observed in modal radiation efficiency curves of modes with higher order axial modal index at medium frequencies. This is attributed to the interference of sound radiated by neighbouring vibrating cells when the distance between nodal lines of a vibrating mode is in the same order or smaller than the acoustic wavelength. Effects of the internal sound field on modal radiation efficiencies of a finite open-end cylinder are discussed.

Content-based video indexing for sports applications using integrated multi-modal approach

To sustain an ongoing rapid growth of video information, there is an emerging demand for a sophisticated content-based video indexing system. However, current video indexing solutions are still immature and lack of any standard. This doctoral consists of a research work based on an integrated multi-modal approach for sports video indexing and retrieval. By combining specific features extractable from multiple audio-visual modalities, generic structure and specific events can be detected and classified. During browsing and retrieval, users will benefit from the integration of high-level semantic and some descriptive mid-level features such as whistle and close-up view of player(s).