Positron emission tomography (PET) tumor segmentation and quantification: Development of new algorithms

Tumor functional volume (FV) and its mean activity concentration (mAC) are the quantities derived from positron emission tomography (PET). These quantities are used for estimating radiation dose for a therapy, evaluating the progression of a disease and also use it as a prognostic indicator for predicting outcome. PET images have low resolution, high noise and affected by partial volume effect (PVE). Manually segmenting each tumor is very cumbersome and very hard to reproduce. To solve the above problem I developed an algorithm, called iterative deconvolution thresholding segmentation (IDTS) algorithm; the algorithm segment the tumor, measures the FV, correct for the PVE and calculates mAC. The algorithm corrects for the PVE without the need to estimate camera's point spread function (PSF); also does not require optimizing for a specific camera. My algorithm was tested in physical phantom studies, where hollow spheres (0.5-16 ml) were used to represent tumors with a homogeneous activity distribution. It was also tested on irregular shaped tumors with a heterogeneous activity profile which were acquired using physical and simulated phantom. The physical phantom studies were performed with different signal to background ratios (SBR) and with different acquisition times (1-5 min). The algorithm was applied on ten clinical data where the results were compared with manual segmentation and fixed percentage thresholding method called T50 and T60 in which 50% and 60% of the maximum intensity respectively is used as threshold. The average error in FV and mAC calculation was 30% and -35% for 0.5 ml tumor. The average error FV and mAC calculation were ~5% for 16 ml tumor. The overall FV error was ∼10% for heterogeneous tumors in physical and simulated phantom data. The FV and mAC error for clinical image compared to manual segmentation was around -17% and 15% respectively. In summary my algorithm has potential to be applied on data acquired from different cameras as its not dependent on knowing the camera's PSF. The algorithm can also improve dose estimation and treatment planning.^

Positron Emission Tomography (PET) Tumor Segmentation and Quantification: Development of New Algorithms

Tumor functional volume (FV) and its mean activity concentration (mAC) are the quantities derived from positron emission tomography (PET). These quantities are used for estimating radiation dose for a therapy, evaluating the progression of a disease and also use it as a prognostic indicator for predicting outcome. PET images have low resolution, high noise and affected by partial volume effect (PVE). Manually segmenting each tumor is very cumbersome and very hard to reproduce. To solve the above problem I developed an algorithm, called iterative deconvolution thresholding segmentation (IDTS) algorithm; the algorithm segment the tumor, measures the FV, correct for the PVE and calculates mAC. The algorithm corrects for the PVE without the need to estimate camera’s point spread function (PSF); also does not require optimizing for a specific camera. My algorithm was tested in physical phantom studies, where hollow spheres (0.5-16 ml) were used to represent tumors with a homogeneous activity distribution. It was also tested on irregular shaped tumors with a heterogeneous activity profile which were acquired using physical and simulated phantom. The physical phantom studies were performed with different signal to background ratios (SBR) and with different acquisition times (1-5 min). The algorithm was applied on ten clinical data where the results were compared with manual segmentation and fixed percentage thresholding method called T50 and T60 in which 50% and 60% of the maximum intensity respectively is used as threshold. The average error in FV and mAC calculation was 30% and -35% for 0.5 ml tumor. The average error FV and mAC calculation were ~5% for 16 ml tumor. The overall FV error was ~10% for heterogeneous tumors in physical and simulated phantom data. The FV and mAC error for clinical image compared to manual segmentation was around -17% and 15% respectively. In summary my algorithm has potential to be applied on data acquired from different cameras as its not dependent on knowing the camera’s PSF. The algorithm can also improve dose estimation and treatment planning.

Automated road pavement marking detection from high resolution aerial images based on Multi-resolution image analysis and anisotropic Gaussian filtering

Road features extraction from remote sensed imagery has been a long-term topic of great interest within the photogrammetry and remote sensing communities for over three decades. The majority of the early work only focused on linear feature detection approaches, with restrictive assumption on image resolution and road appearance. The widely available of high resolution digital aerial images makes it possible to extract sub-road features, e.g. road pavement markings. In this paper, we will focus on the automatic extraction of road lane markings, which are required by various lane-based vehicle applications, such as, autonomous vehicle navigation, and lane departure warning. The proposed approach consists of three phases: i) road centerline extraction from low resolution image, ii) road surface detection in the original image, and iii) pavement marking extraction on the generated road surface. The proposed method was tested on the aerial imagery dataset of the Bruce Highway, Queensland, and the results demonstrate the efficiency of our approach.

Automatic tracking, super-resolution and recognition of human faces from surveillance video

Identifying an individual from surveillance video is a difficult, time consuming and labour intensive process. The proposed system aims to streamline this process by filtering out unwanted scenes and enhancing an individual's face through super-resolution. An automatic face recognition system is then used to identify the subject or present the human operator with likely matches from a database. A person tracker is used to speed up the subject detection and super-resolution process by tracking moving subjects and cropping a region of interest around the subject's face to reduce the number and size of the image frames to be super-resolved respectively. In this paper, experiments have been conducted to demonstrate how the optical flow super-resolution method used improves surveillance imagery for visual inspection as well as automatic face recognition on an Eigenface and Elastic Bunch Graph Matching system. The optical flow based method has also been benchmarked against the ``hallucination'' algorithm, interpolation methods and the original low-resolution images. Results show that both super-resolution algorithms improved recognition rates significantly. Although the hallucination method resulted in slightly higher recognition rates, the optical flow method produced less artifacts and more visually correct images suitable for human consumption.

Focus-score weighted super-resolution for uncooperative iris recognition at a distance and on the move

Uncooperative iris identification systems at a distance and on the move often suffer from poor resolution and poor focus of the captured iris images. The lack of pixel resolution and well-focused images significantly degrades the iris recognition performance. This paper proposes a new approach to incorporate the focus score into a reconstruction-based super-resolution process to generate a high resolution iris image from a low resolution and focus inconsistent video sequence of an eye. A reconstruction-based technique, which can incorporate middle and high frequency components from multiple low resolution frames into one desired super-resolved frame without introducing false high frequency components, is used. A new focus assessment approach is proposed for uncooperative iris at a distance and on the move to improve performance for variations in lighting, size and occlusion. A novel fusion scheme is then proposed to incorporate the proposed focus score into the super-resolution process. The experiments conducted on the The Multiple Biometric Grand Challenge portal database shows that our proposed approach achieves an EER of 2.1%, outperforming the existing state-of-the-art averaging signal-level fusion approach by 19.2% and the robust mean super-resolution approach by 8.7%.

Feature-domain super-resolution framework for gabor-based face and iris recognition

The low resolution of images has been one of the major limitations in recognising humans from a distance using their biometric traits, such as face and iris. Superresolution has been employed to improve the resolution and the recognition performance simultaneously, however the majority of techniques employed operate in the pixel domain, such that the biometric feature vectors are extracted from a super-resolved input image. Feature-domain superresolution has been proposed for face and iris, and is shown to further improve recognition performance by capitalising on direct super-resolving the features which are used for recognition. However, current feature-domain superresolution approaches are limited to simple linear features such as Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA), which are not the most discriminant features for biometrics. Gabor-based features have been shown to be one of the most discriminant features for biometrics including face and iris. This paper proposes a framework to conduct super-resolution in the non-linear Gabor feature domain to further improve the recognition performance of biometric systems. Experiments have confirmed the validity of the proposed approach, demonstrating superior performance to existing linear approaches for both face and iris biometrics.

Improved Subject Identification in Surveillance Video using Super Resolution

The time consuming and labour intensive task of identifying individuals in surveillance video is often challenged by poor resolution and the sheer volume of stored video. Faces or identifying marks such as tattoos are often too coarse for direct matching by machine or human vision. Object tracking and super-resolution can then be combined to facilitate the automated detection and enhancement of areas of interest. The object tracking process enables the automatic detection of people of interest, greatly reducing the amount of data for super-resolution. Smaller regions such as faces can also be tracked. A number of instances of such regions can then be utilized to obtain a super-resolved version for matching. Performance improvement from super-resolution is demonstrated using a face verification task. It is shown that there is a consistent improvement of approximately 7% in verification accuracy, using both Eigenface and Elastic Bunch Graph Matching approaches for automatic face verification, starting from faces with an eye to eye distance of 14 pixels. Visual improvement in image fidelity from super-resolved images over low-resolution and interpolated images is demonstrated on a small database. Current research and future directions in this area are also summarized.

Vision-based place recognition : how low can you go?

In this paper we use the algorithm SeqSLAM to address the question, how little and what quality of visual information is needed to localize along a familiar route? We conduct a comprehensive investigation of place recognition performance on seven datasets while varying image resolution (primarily 1 to 512 pixel images), pixel bit depth, field of view, motion blur, image compression and matching sequence length. Results confirm that place recognition using single images or short image sequences is poor, but improves to match or exceed current benchmarks as the matching sequence length increases. We then present place recognition results from two experiments where low-quality imagery is directly caused by sensor limitations; in one, place recognition is achieved along an unlit mountain road by using noisy, long-exposure blurred images, and in the other, two single pixel light sensors are used to localize in an indoor environment. We also show failure modes caused by pose variance and sequence aliasing, and discuss ways in which they may be overcome. By showing how place recognition along a route is feasible even with severely degraded image sequences, we hope to provoke a re-examination of how we develop and test future localization and mapping systems.

Higher-resolution structure of the human insulin receptor ectodomain: Multi-modal inclusion of the insert domain

Insulin receptor (IR) signaling is critical to controlling nutrient uptake and metabolism. However, only a low-resolution (3.8 Å) structure currently exists for the IR ectodomain, with some segments ill-defined or unmodeled due to disorder. Here, we revise this structure using new diffraction data to 3.3 Å resolution that allow improved modeling of the N-linked glycans, the first and third fibronectin type III domains, and the insert domain. A novel haptic interactive molecular dynamics strategy was used to aid fitting to low-resolution electron density maps. The resulting model provides a foundation for investigation of structural transitions in IR upon ligand binding.

Structure of Sesbania Mosaic Virus at 4·7 Å Resolution and Partial Sequence of the Coat Protein

Sesbania mosaic virus (SMV) is a plant virus infecting Sesbania grandiflora plants in Andhra Pradesh, India. Amino acid sequence of the tryptic peptides of SMV coat protein were determined using a gas phase sequenator. These sequences showed identical amino acids at 69% of the positions when aligned with the corresponding residues of southern bean mosaic virus (SBMV).Crystals diffracting to better than 3 Å resolution were obtained by precipitating the virus with ammonium sulphate. The crystals belonged to rhombohedral space group R3 with α = 291·4 Å and α = 61·9°. Three-dimensional X-ray diffraction data on these crystals were collected to a resolution of 4·7 Å, using a Siemens-Nicolet area detector system. Self-rotation function studies revealed the icosahedral symmetry of the virus particles, as well as their precise orientation in the unit cell. Cross-rotation function and modelling studies with SBMV showed that it is a valid starting model for SMV structure determination. Low resolution phases computed using a polyalanine model of SBMV were subjected to refinement and extension by real-space electron density averaging and solvent flattening. The final electron density map revealed a polypeptide fold similar to SBMV. The single disulphide bridge of SBMV coat protein is retained in SMV. Four icosahedrally independent cation binding sites have been tentatively identified. Three of these sites, related by a quasi threefold axis, are also found in SBMV. The fourth site is situated on the quasi threefold axis. Aspartic acid residues, which replace Ile218 of SBMV from the quasi threefold-related subunits are suitable ligands to the cation at this site

HIGH-RESOLUTION IMAGING OF THE ATLBS REGIONS: THE RADIO SOURCE COUNTS

The Australia Telescope Low-brightness Survey (ATLBS) regions have been mosaic imaged at a radio frequency of 1.4 GHz with 6 `' angular resolution and 72 mu Jy beam(-1) rms noise. The images (centered at R. A. 00(h)35(m)00(s), decl. -67 degrees 00'00 `' and R. A. 00(h)59(m)17(s), decl. -67.00'00 `', J2000 epoch) cover 8.42 deg(2) sky area and have no artifacts or imaging errors above the image thermal noise. Multi-resolution radio and optical r-band images (made using the 4 m CTIO Blanco telescope) were used to recognize multi-component sources and prepare a source list; the detection threshold was 0.38 mJy in a low-resolution radio image made with beam FWHM of 50 `'. Radio source counts in the flux density range 0.4-8.7 mJy are estimated, with corrections applied for noise bias, effective area correction, and resolution bias. The resolution bias is mitigated using low-resolution radio images, while effects of source confusion are removed by using high-resolution images for identifying blended sources. Below 1 mJy the ATLBS counts are systematically lower than the previous estimates. Showing no evidence for an upturn down to 0.4 mJy, they do not require any changes in the radio source population down to the limit of the survey. The work suggests that automated image analysis for counts may be dependent on the ability of the imaging to reproduce connecting emission with low surface brightness and on the ability of the algorithm to recognize sources, which may require that source finding algorithms effectively work with multi-resolution and multi-wavelength data. The work underscores the importance of using source lists-as opposed to component lists-and correcting for the noise bias in order to precisely estimate counts close to the image noise and determine the upturn at sub-mJy flux density.

The Relationship Between Solar Coronal X-Ray Brightness and Active Region Magnetic Fields: A Study Using High-Resolution Hinode Observations

By using high-resolution observations of nearly co-temporal and co-spatial Solar Optical Telescope spectropolarimeter and X-Ray Telescope coronal X-ray data onboard Hinode, we revisit the problematic relationship between global magnetic quantities and coronal X-ray brightness. Co-aligned vector magnetogram and X-ray data were used for this study. The total X-ray brightness over active regions is well correlated with integrated magnetic quantities such as the total unsigned magnetic flux, the total unsigned vertical current, and the area-integrated square of the vertical and horizontal magnetic fields. On accounting for the inter-dependence of the magnetic quantities, we inferred that the total magnetic flux is the primary determinant of the observed integrated X-ray brightness. Our observations indicate that a stronger coronal X-ray flux is not related to a higher non-potentiality of active-region magnetic fields. The data even suggest a slightly negative correlation between X-ray brightness and a proxy of active-region non-potentiality. Although there are small numerical differences in the established correlations, the main conclusions are qualitatively consistent over two different X-ray filters, the Al-poly and Ti-poly filters, which confirms the strength of our conclusions and validate and extend earlier studies that used low-resolution data. We discuss the implications of our results and the constraints they set on theories of solar coronal heating.

Super Resolution Using a Single Image Dictionary

To perform super resolution of low resolution images, state-of-the-art methods are based on learning a pair of lowresolution and high-resolution dictionaries from multiple images. These trained dictionaries are used to replace patches in lowresolution image with appropriate matching patches from the high-resolution dictionary. In this paper we propose using a single common image as dictionary, in conjunction with approximate nearest neighbour fields (ANNF) to perform super resolution (SR). By using a common source image, we are able to bypass the learning phase and also able to reduce the dictionary from a collection of hundreds of images to a single image. By adapting recent developments in ANNF computation, to suit super-resolution, we are able to perform much faster and accurate SR than existing techniques. To establish this claim, we compare the proposed algorithm against various state-of-the-art algorithms, and show that we are able to achieve b etter and faster reconstruction without any training.

Synthetic image super resolution using FeatureMatch

In this paper, we propose a super resolution (SR) method for synthetic images using FeatureMatch. Existing state-of-the-art super resolution methods are learning based methods, where a pair of low-resolution and high-resolution dictionary pair are trained, and this trained pair is used to replace patches in low-resolution image with appropriate matching patches from the high-resolution dictionary. In this paper, we show that by using Approximate Nearest Neighbour Fields (ANNF), and a common source image, we can by-pass the learning phase, and use a single image for dictionary. Thus, reducing the dictionary from a collection obtained from hundreds of training images, to a single image. We show that by modifying the latest developments in ANNF computation, to suit super resolution, we can perform much faster and more accurate SR than existing techniques. To establish this claim we will compare our algorithm against various state-of-the-art algorithms, and show that we are able to achieve better and faster reconstruction without any training phase.

Joint modeling of lithosphere and mantle dynamics: Evaluation of constraints from global tomography models

With the advances in technology, seismological theory, and data acquisition, a number of high-resolution seismic tomography models have been published. However, discrepancies between tomography models often arise from different theoretical treatments of seismic wave propagation, different inversion strategies, and different data sets. Using a fixed velocity-to-density scaling and a fixed radial viscosity profile, we compute global mantle flow models associated with the different tomography models and test the impact of these for explaining surface geophysical observations (geoid, dynamic topography, stress, and strain rates). We use the joint modeling of lithosphere and mantle dynamics approach of Ghosh and Holt (2012) to compute the full lithosphere stresses, except that we use HC for the mantle circulation model, which accounts for the primary flow-coupling features associated with density-driven mantle flow. Our results show that the seismic tomography models of S40RTS and SAW642AN provide a better match with surface observables on a global scale than other models tested. Both of these tomography models have important similarities, including upwellings located in Pacific, Eastern Africa, Iceland, and mid-ocean ridges in the Atlantic and Indian Ocean and downwelling flows mainly located beneath the Andes, the Middle East, and central and Southeast Asia.