Ontology-based image annotation

With regard to the long-standing problem of the semantic gap between low-level image features and high-level human knowledge, the image retrieval community has recently shifted its emphasis from low-level features analysis to high-level image semantics extrac- tion. User studies reveal that users tend to seek information using high-level semantics. Therefore, image semantics extraction is of great importance to content-based image retrieval because it allows the users to freely express what images they want. Semantic content annotation is the basis for semantic content retrieval. The aim of image anno- tation is to automatically obtain keywords that can be used to represent the content of images. The major research challenges in image semantic annotation are: what is the basic unit of semantic representation? how can the semantic unit be linked to high-level image knowledge? how can the contextual information be stored and utilized for image annotation? In this thesis, the Semantic Web technology (i.e. ontology) is introduced to the image semantic annotation problem. Semantic Web, the next generation web, aims at mak- ing the content of whatever type of media not only understandable to humans but also to machines. Due to the large amounts of multimedia data prevalent on the Web, re- searchers and industries are beginning to pay more attention to the Multimedia Semantic Web. The Semantic Web technology provides a new opportunity for multimedia-based applications, but the research in this area is still in its infancy. Whether ontology can be used to improve image annotation and how to best use ontology in semantic repre- sentation and extraction is still a worth-while investigation. This thesis deals with the problem of image semantic annotation using ontology and machine learning techniques in four phases as below. 1) Salient object extraction. A salient object servers as the basic unit in image semantic extraction as it captures the common visual property of the objects. Image segmen- tation is often used as the �rst step for detecting salient objects, but most segmenta- tion algorithms often fail to generate meaningful regions due to over-segmentation and under-segmentation. We develop a new salient object detection algorithm by combining multiple homogeneity criteria in a region merging framework. 2) Ontology construction. Since real-world objects tend to exist in a context within their environment, contextual information has been increasingly used for improving object recognition. In the ontology construction phase, visual-contextual ontologies are built from a large set of fully segmented and annotated images. The ontologies are composed of several types of concepts (i.e. mid-level and high-level concepts), and domain contextual knowledge. The visual-contextual ontologies stand as a user-friendly interface between low-level features and high-level concepts. 3) Image objects annotation. In this phase, each object is labelled with a mid-level concept in ontologies. First, a set of candidate labels are obtained by training Support Vectors Machines with features extracted from salient objects. After that, contextual knowledge contained in ontologies is used to obtain the �nal labels by removing the ambiguity concepts. 4) Scene semantic annotation. The scene semantic extraction phase is to get the scene type by using both mid-level concepts and domain contextual knowledge in ontologies. Domain contextual knowledge is used to create scene con�guration that describes which objects co-exist with which scene type more frequently. The scene con�guration is represented in a probabilistic graph model, and probabilistic inference is employed to calculate the scene type given an annotated image. To evaluate the proposed methods, a series of experiments have been conducted in a large set of fully annotated outdoor scene images. These include a subset of the Corel database, a subset of the LabelMe dataset, the evaluation dataset of localized semantics in images, the spatial context evaluation dataset, and the segmented and annotated IAPR TC-12 benchmark.

A unifying view of multiple kernel learning

Recent research on multiple kernel learning has lead to a number of approaches for combining kernels in regularized risk minimization. The proposed approaches include different formulations of objectives and varying regularization strategies. In this paper we present a unifying optimization criterion for multiple kernel learning and show how existing formulations are subsumed as special cases. We also derive the criterion’s dual representation, which is suitable for general smooth optimization algorithms. Finally, we evaluate multiple kernel learning in this framework analytically using a Rademacher complexity bound on the generalization error and empirically in a set of experiments.

Improved image quality for x-ray CT imaging of gel dosimeters

Purpose: This study provides a simple method for improving precision of x-ray computed tomography (CT) scans of irradiated polymer gel dosimetry. The noise affecting CT scans of irradiated gels has been an impediment to the use of clinical CT scanners for gel dosimetry studies. Method: In this study, it is shown that multiple scans of a single PAGAT gel dosimeter can be used to extrapolate a ‘zero-scan’ image which displays a similar level of precision to an image obtained by averaging multiple CT images, without the compromised dose measurement resulting from the exposure of the gel to radiation from the CT scanner. Results: When extrapolating the zero-scan image, it is shown that exponential and simple linear fits to the relationship between Hounsfield unit and scan number, for each pixel in the image, provides an accurate indication of gel density. Conclusions: It is expected that this work will be utilised in the analysis of three-dimensional gel volumes irradiated using complex radiotherapy treatments.

Implementation of a survey form marking system using image processing techniques on an Imputer

This paper describes the feasibility of the application of an Imputer in a multiple choice answer sheet marking system based on image processing techniques.

Speckle reduction in optical coherence tomography imaging by affine-motion image registration

Signal-degrading speckle is one factor that can reduce the quality of optical coherence tomography images. We demonstrate the use of a hierarchical model-based motion estimation processing scheme based on an affine-motion model to reduce speckle in optical coherence tomography imaging, by image registration and the averaging of multiple B-scans. The proposed technique is evaluated against other methods available in the literature. The results from a set of retinal images show the benefit of the proposed technique, which provides an improvement in signal-to-noise ratio of the square root of the number of averaged images, leading to clearer visual information in the averaged image. The benefits of the proposed technique are also explored in the case of ocular anterior segment imaging.

Speaker attribution of multiple telephone conversations using a complete-linkage clustering approach

In this paper we propose and evaluate a speaker attribution system using a complete-linkage clustering method. Speaker attribution refers to the annotation of a collection of spoken audio based on speaker identities. This can be achieved using diarization and speaker linking. The main challenge associated with attribution is achieving computational efficiency when dealing with large audio archives. Traditional agglomerative clustering methods with model merging and retraining are not feasible for this purpose. This has motivated the use of linkage clustering methods without retraining. We first propose a diarization system using complete-linkage clustering and show that it outperforms traditional agglomerative and single-linkage clustering based diarization systems with a relative improvement of 40% and 68%, respectively. We then propose a complete-linkage speaker linking system to achieve attribution and demonstrate a 26% relative improvement in attribution error rate (AER) over the single-linkage speaker linking approach.

Combining multiple sensor modalities for a localisation robust to smoke

This paper proposes an approach to obtain a localisation that is robust to smoke by exploiting multiple sensing modalities: visual and infrared (IR) cameras. This localisation is based on a state-of-the-art visual SLAM algorithm. First, we show that a reasonably accurate localisation can be obtained in the presence of smoke by using only an IR camera, a sensor that is hardly affected by smoke, contrary to a visual camera (operating in the visible spectrum). Second, we demonstrate that improved results can be obtained by combining the information from the two sensor modalities (visual and IR cameras). Third, we show that by detecting the impact of smoke on the visual images using a data quality metric, we can anticipate and mitigate the degradation in performance of the localisation by discarding the most affected data. The experimental validation presents multiple trajectories estimated by the various methods considered, all thoroughly compared to an accurate dGPS/INS reference.

Identifying steganographic payload location in binary image

In this paper, we propose a steganalysis method that is able to identify the locations of stego bearing pixels in the binary image. In order to do that, our proposed method will calculate the residual between a given stego image and its estimated cover image. After that, we will compute the local entropy difference between these two versions of images as well. Finally, we will compute the mean of residual and mean of local entropy difference across multiple stego images. From these two means, the locations of stego bearing pixels can be identified. The presented empirical results demonstrate that our proposed method can identify the stego bearing locations of near perfect accuracy when sufficient stego images are supplied. Hence, our proposed method can be used to reveal which pixels in the binary image have been used to carry the secret message.

Matching image sets via adaptive multi convex hull

Traditional nearest points methods use all the samples in an image set to construct a single convex or affine hull model for classification. However, strong artificial features and noisy data may be generated from combinations of training samples when significant intra-class variations and/or noise occur in the image set. Existing multi-model approaches extract local models by clustering each image set individually only once, with fixed clusters used for matching with various image sets. This may not be optimal for discrimination, as undesirable environmental conditions (eg. illumination and pose variations) may result in the two closest clusters representing different characteristics of an object (eg. frontal face being compared to non-frontal face). To address the above problem, we propose a novel approach to enhance nearest points based methods by integrating affine/convex hull classification with an adapted multi-model approach. We first extract multiple local convex hulls from a query image set via maximum margin clustering to diminish the artificial variations and constrain the noise in local convex hulls. We then propose adaptive reference clustering (ARC) to constrain the clustering of each gallery image set by forcing the clusters to have resemblance to the clusters in the query image set. By applying ARC, noisy clusters in the query set can be discarded. Experiments on Honda, MoBo and ETH-80 datasets show that the proposed method outperforms single model approaches and other recent techniques, such as Sparse Approximated Nearest Points, Mutual Subspace Method and Manifold Discriminant Analysis.

Scleral matrix metalloproteinases, serine proteinase activity and hydrational capacity are increased in myopia induced by retinal image degradation

In the avian model of myopia, retinal image degradation quickly leads to ocular enlargement. We now give evidence that regionally specific changes in ocular size are correlated with both biomechanical indices of scleral remodeling, e.g. hydration capacity and with biochemical changes in proteinase activities. The latter include a 72 kDa matrix metalloproteinase (putatively MMP-2), other gelatin-binding MMPs, an acid pH MMP and a serine protease. Specifically, we have found that increases in scleral hydrational capacity parallel increases in collagen degrading activities. Gelatin zymography reveals that eyes with 7 days of retinal image degradation have elevated levels (1.4-fold) of gelatinolytic activities at 72 and 67 kDa M(r) in equatorial and posterior pole regions of the sclera while, after 14 days of treatment, increases are no longer apparent. Lower M(r) zymographic activities at 50, 46 and 37 kDa M(r) are collectively increased in eyes treated for both 7 and 14 days (1.4- and 2.4-fold respectively) in the equator and posterior pole areas of enlarging eyes. Western blot analyses of scleral extracts with an antibody to human MMP-2 reveals immunoreactive bands at 65, 30 and 25 kDa. Zymograms incubated under slightly acidic conditions reveal that, in enlarging eyes, MMP activities at 25 and 28 kDa M(r) are increased in scleral equator and posterior pole (1.6- and 4.5-fold respectively). A TIMP-like protein is also identified in sclera and cornea by Western blot analysis. Finally, retinal-image degradation also increases (~2.6-fold) the activity of a 23.5 kDa serine proteinase in limbus, equator and posterior pole sclera that is inhibited by aprotinin and soybean trypsin inhibitor. Taken together, these results indicate that eye growth induced by retinal-image degradation involves increases in the activities of multiple scleral proteinases that could modify the biomechanical properties of scleral structural components and contribute to tissue remodeling and growth.

Robust multiple-sensing-modality data fusion using Gaussian Process Implicit Surfaces

The ability to build high-fidelity 3D representations of the environment from sensor data is critical for autonomous robots. Multi-sensor data fusion allows for more complete and accurate representations. Furthermore, using distinct sensing modalities (i.e. sensors using a different physical process and/or operating at different electromagnetic frequencies) usually leads to more reliable perception, especially in challenging environments, as modalities may complement each other. However, they may react differently to certain materials or environmental conditions, leading to catastrophic fusion. In this paper, we propose a new method to reliably fuse data from multiple sensing modalities, including in situations where they detect different targets. We first compute distinct continuous surface representations for each sensing modality, with uncertainty, using Gaussian Process Implicit Surfaces (GPIS). Second, we perform a local consistency test between these representations, to separate consistent data (i.e. data corresponding to the detection of the same target by the sensors) from inconsistent data. The consistent data can then be fused together, using another GPIS process, and the rest of the data can be combined as appropriate. The approach is first validated using synthetic data. We then demonstrate its benefit using a mobile robot, equipped with a laser scanner and a radar, which operates in an outdoor environment in the presence of large clouds of airborne dust and smoke.

Simultaneous localization and planning on multiple map hypotheses

This paper presents a novel method to rank map hypotheses by the quality of localization they afford. The highest ranked hypothesis at any moment becomes the active representation that is used to guide the robot to its goal location. A single static representation is insufficient for navigation in dynamic environments where paths can be blocked periodically, a common scenario which poses significant challenges for typical planners. In our approach we simultaneously rank multiple map hypotheses by the influence that localization in each of them has on locally accurate odometry. This is done online for the current locally accurate window by formulating a factor graph of odometry relaxed by localization constraints. Comparison of the resulting perturbed odometry of each hypothesis with the original odometry yields a score that can be used to rank map hypotheses by their utility. We deploy the proposed approach on a real robot navigating a structurally noisy office environment. The configuration of the environment is physically altered outside the robots sensory horizon during navigation tasks to demonstrate the proposed approach of hypothesis selection.

Data fusion and tracking with multiple UAVs

This chapter describes decentralized data fusion algorithms for a team of multiple autonomous platforms. Decentralized data fusion (DDF) provides a useful basis with which to build upon for cooperative information gathering tasks for robotic teams operating in outdoor environments. Through the DDF algorithms, each platform can maintain a consistent global solution from which decisions may then be made. Comparisons will be made between the implementation of DDF using two probabilistic representations. The first, Gaussian estimates and the second Gaussian mixtures are compared using a common data set. The overall system design is detailed, providing insight into the overall complexity of implementing a robust DDF system for use in information gathering tasks in outdoor UAV applications.

Automatic population HARDI white matter tract clustering by label fusion of multiple tract atlases

Automatic labeling of white matter fibres in diffusion-weighted brain MRI is vital for comparing brain integrity and connectivity across populations, but is challenging. Whole brain tractography generates a vast set of fibres throughout the brain, but it is hard to cluster them into anatomically meaningful tracts, due to wide individual variations in the trajectory and shape of white matter pathways. We propose a novel automatic tract labeling algorithm that fuses information from tractography and multiple hand-labeled fibre tract atlases. As streamline tractography can generate a large number of false positive fibres, we developed a top-down approach to extract tracts consistent with known anatomy, based on a distance metric to multiple hand-labeled atlases. Clustering results from different atlases were fused, using a multi-stage fusion scheme. Our "label fusion" method reliably extracted the major tracts from 105-gradient HARDI scans of 100 young normal adults. © 2012 Springer-Verlag.

Birdcall retrieval from environmental acoustic recordings using image processing

Acoustic recordings of the environment provide an effective means to monitor bird species diversity. To facilitate exploration of acoustic recordings, we describe a content-based birdcall retrieval algorithm. A query birdcall is a region of spectrogram bounded by frequency and time. Retrieval depends on a similarity measure derived from the orientation and distribution of spectral ridges. The spectral ridge detection method caters for a broad range of birdcall structures. In this paper, we extend previous work by incorporating a spectrogram scaling step in order to improve the detection of spectral ridges. Compared to an existing approach based on MFCC features, our feature representation achieves better retrieval performance for multiple bird species in noisy recordings.