Image Analysts Research Articles

In vivo quantification of amygdala subnuclei using 4.7 T fast spin echo imaging

The amygdala (AG) is an almond-shaped heterogeneous structure located in the medial temporal lobe. The majority of previous structural Magnetic Resonance Imaging (MRI) volumetric methods for AG measurement have so far only been able to examine this region as a whole. In order to understand the role of the AG in different neuropsychiatric disorders, it is necessary to understand the functional role of its subnuclei. The main goal of the present study was to develop a reliable volumetric method to delineate major AG subnuclei groups using ultra-high resolution high field MRI.38 healthy volunteers (15 males and 23 females, 21–60 years of age) without any history of medical or neuropsychiatric disorders were recruited for this study. Structural MRI datasets were acquired at 4.7 T Varian Inova MRI system using a fast spin echo (FSE) sequence.The AG was manually segmented into its five major anatomical subdivisions: lateral (La), basal (B), accessory basal (AB) nuclei, and cortical (Co) and centromedial (CeM) groups. Inter-(intra-) rater reliability of our novel volumetric method was assessed using intra-class correlation coefficient (ICC) and Dice's Kappa.Our results suggest that reliable measurements of the AG subnuclei can be obtained by image analysts with experience in AG anatomy. We provided a step-by-step segmentation protocol and reported absolute and relative volumes for the AG subnuclei. Our results showed that the basolateral (BLA) complex occupies seventy-eight percent of the total AG volume, while CeM and Co groups occupy twenty-two percent of the total AG volume. Finally, we observed no hemispheric effects and no gender differences in the total AG volume and the volumes of its subnuclei.Future applications of this method will help to understand the selective vulnerability of the AG subnuclei in neurological and psychiatric disorders.

Workshop on reconstruction schemes for magnetic resonance data: summary of findings and recommendations.

The high-fidelity reconstruction of compressed and low-resolution magnetic resonance (MR) data is essential for simultaneously improving patient care, accuracy in diagnosis and quality in clinical research. Sponsored by the Royal Society through the Newton Mobility Grant Scheme, we held a half-day workshop on reconstruction schemes for MR data on 17 August 2016 to discuss new ideas from related research fields that could be useful to overcome the shortcomings of the conventional reconstruction methods that have been evaluated to date. Participants were 21 university students, computer scientists, image analysts, engineers and physicists from institutions from six different countries. The discussion evolved around exploring new avenues to achieve high resolution, high quality and fast acquisition of MR imaging. In this article, we summarize the topics covered throughout the workshop and make recommendations for ongoing and future works.

Bone marrow fat content is correlated with hepatic fat content in paediatric non-alcoholic fatty liver disease

To investigate the relationship between bone marrow fat content and hepatic fat content in children with known or suspected non-alcoholic fatty liver disease (NAFLD). This was an institutional review board-approved, Health Insurance Portability and Accountability Act (HIPAA)-compliant, cross-sectional, prospective analysis of data collected between October 2010 to March 2013 in 125 children with known or suspected NAFLD. Written informed consent was obtained for same-day research magnetic resonance imaging (MRI) of the lumbar spine, liver, and abdominal adiposity. Lumbar spine bone marrow proton density fat fraction (PDFF) and hepatic PDFF were estimated using complex-based MRI (C-MRI) techniques and magnitude-based MRI (M-MRI), respectively. Visceral adipose tissue (VAT) and subcutaneous adipose tissue (SCAT) were quantified using high-resolution MRI. All images were acquired by two MRI technologists. Hepatic M-MRI images were analysed by an image analyst; all other images were analysed by a single investigator. The relationship between lumbar spine bone marrow PDFF and hepatic PDFF was assessed with and without adjusting for the presence of covariates using correlation and regression analysis. Lumbar spine bone marrow PDFF was positively associated with hepatic PDFF in children with known or suspected NAFLD prior to adjusting for covariates (r=0.33, p=0.0002). Lumbar spine bone marrow PDFF was positively associated with hepatic PDFF in children with known or suspected NAFLD (r=0.24, p=0.0079) after adjusting for age, sex, body mass index z-score, VAT, and SCAT in a multivariable regression analysis. Bone marrow fat content is positively associated with hepatic fat content in children with known or suspected NAFLD. Further research is needed to confirm these results and understand their clinical and biological implications.

Developing the Quantitative Histopathology Image Ontology (QHIO): A case study using the hot spot detection problem

Interoperability across data sets is a key challenge for quantitative histopathological imaging. There is a need for an ontology that can support effective merging of pathological image data with associated clinical and demographic data. To foster organized, cross-disciplinary, information-driven collaborations in the pathological imaging field, we propose to develop an ontology to represent imaging data and methods used in pathological imaging and analysis, and call it Quantitative Histopathological Imaging Ontology – QHIO. We apply QHIO to breast cancer hot-spot detection with the goal of enhancing reliability of detection by promoting the sharing of data between image analysts.

Semi-Automatic Selection of Ground Control Points for High Resolution Remote Sensing Data in Urban Areas

Abstract Geometrical accuracy of remote sensing data often is ensured by geometrical transforms based on Ground Control Points (GCPs). Manual selection of GCP is a time-consuming process, which requires some sort of automation. Therefore, the aim of this study is to present and evaluate methodology for easier, semi-automatic selection of ground control points for urban areas. Custom line scanning algorithm was implemented and applied to data in order to extract potential GCPs for an image analyst. The proposed method was tested for classical orthorectification and special object polygon transform. Results are convincing and show that in the test case semi-automatic methodology is able to correct locations of 70 % (thermal data) – 80 % (orthophoto images) of buildings. Geometrical transform for subimages of approximately 3 hectares with approximately 12 automatically found GCPs resulted in RSME approximately 1 meter with standard deviation of 1.2 meters.

Trends and perspectives of space-borne SAR remote sensing for archaeological landscape and cultural heritage applications

This paper provides an overview of the opportunities that image analysts, archaeologists and conservation scientists currently have to use space-borne Synthetic Aperture Radar (SAR) imagery for prospection of cultural landscapes and investigation of environmental, land surface and anthropogenic processes that can alter the condition of heritage assets. The benefits of the recent developments in SAR satellite sensors towards higher resolution (up to less than 1m) and shorter revisiting times (up to a few days) are discussed in relation to established techniques using the two key SAR parameters – amplitude and phase. Selected case studies from Middle East to South America illustrate how SAR can be effectively used to detect subtle archaeological features in modern landscapes, monitor historic sites and assess damage in areas of conflict. These examples form the basis to highlight the current trends in archaeological remote sensing based on space-borne SAR data in the era of the European Space Agency's Sentinel-1 constellation and on-demand high resolution space missions such as TerraSAR-X.

VIIRS Day/Night Band—Correcting Striping and Nonuniformity over a Very Large Dynamic Range

The Suomi National Polar-orbiting (NPP) Visible Infrared Imaging Radiometer Suite (VIIRS) Day-Night Band (DNB) measures visible and near-infrared light extending over seven orders of magnitude of dynamic range. This makes radiometric calibration difficult. We have observed that DNB imagery has striping, banding and other nonuniformities—day or night. We identified the causes as stray light, nonlinearity, detector crosstalk, hysteresis and mirror-side variation. We found that these affect both Earth-view and calibration signals. These present an obstacle to interpretation by users of DNB products. Because of the nonlinearity we chose the histogram matching destriping technique which we found is successful for daytime, twilight and nighttime scenes. Because of the very large dynamic range of the DNB, we needed to add special processes to the histogram matching to destripe all scenes, especially imagery in the twilight regions where scene illumination changes rapidly over short distances. We show that destriping aids image analysts, and makes it possible for advanced automated cloud typing algorithms. Manual or automatic identification of other features, including polar ice and gravity waves in the upper atmosphere are also discussed. In consideration of the large volume of data produced 24 h a day by the VIIRS DNB, we present methods for reducing processing time.

‘Looting marks’ in space-borne SAR imagery: Measuring rates of archaeological looting in Apamea (Syria) with TerraSAR-X Staring Spotlight

In archaeological remote sensing, space-borne Synthetic Aperture Radar (SAR) has not been used so far to monitor ‘looting’ (i.e. illegal excavations in heritage sites) mainly because of the spatial resolution of SAR images, typically not comparable to the ground dimensions of looting features. This paper explores the potential of the new TerraSAR-X beam mode Staring Spotlight (ST) to investigate looting within a workflow of radar backscattering change detection. A bespoke time series of five single polarisation, ascending mode, ST scenes with an unprecedented azimuth resolution of 0.24 m was acquired over the archaeological site of Apamea in western Syria, from October 2014 to June 2015 with a regular sampling of one image every two months. Formerly included in the Tentative List of UNESCO, the site has been heavily looted from at least early 2012 to May 2014, as confirmed by Google Earth Very High Resolution (VHR) optical imagery. Building upon the theory of SAR imaging, we develop a novel conceptual model of ‘looting marks’, identify marks due to occurrence of new looting and discriminate them from alteration (e.g. filling) of pre-existing looting holes. ‘Looting marks’ appear as distinctive patterns of shadow and layover which are visible in the ground-range reprojected ST image and generated by the morphology of the holes. The recognition of looting marks within ratio maps of radar backscatter (σ0) between consecutive ST scenes allows quantification of the magnitude, spatial distribution and rates of looting activities. In agreement with the estimates based on Google Earth imagery, the ST acquired in October 2014 shows that ~ 45% of the site was looted. In the following eight months new looting happened locally, with holes mainly dug along the margins of the already looted areas. Texture values of ~ 0.31 clearly distinguish these holes from the unaltered, bare ground nearby. Hot spots of change are identified based on the temporal variability of σ0, and colour composites indicate where repeated looting and alteration of existing holes occurred. Most looting marks are observed north of the two main Roman decumani. Looting intensified almost steadily from December 2014, with over 1500 new marks in February–April 2015. The estimated rates of looting increased from 214 looting marks/month in October–December 2014 to over 780 marks/month in April–June 2015, and numerically express the dynamic nature of the phenomenon to which Apamea is still exposed. The method of identifying looting marks in VHR radar images therefore proves a reliable opportunity for archaeologists and image analysts to measure remotely the scale of looting and monitor its temporal evolution.

English

Associative procedures for content-based image position by semantics are gorgeous due to the correspondence of produced models to human models of considerate. Though they tend to return results that are better unstated by image analysts, the induction of these models is problematic to build due to factors that affect training density, such as synchronicity of visual patterns in same images, over-fitting or under-fitting and semantic representation differences among image analysts. This article proposes a methodology to reduce the complexity of ranking satellite images for associative methods. Our method employs genetic operations to afford faster and more correct models for ranking by semantic using low level features. The added exactness is provided by a decrease in the likelihood to reach local minima or to over fit. The investigates show that, using genetic optimization, associative methods perform better or at similar levels as state-of-the-art ensemble methods for ranking. The mean average precision (MAP) of ranking by semantic was improved by 14% over similar associative methods that use other optimization procedures while maintaining smaller size for each semantic model.

Validation of a Semiautomated Liver Segmentation Method Using CT for Accurate Volumetry

To compare the repeatability and agreement of a semiautomated liver segmentation method with manual segmentation for assessment of total liver volume on CT (computed tomography). This retrospective, institutional review board-approved study was conducted in 41 subjects who underwent liver CT for preoperative planning. The major pathologies encountered were colorectal cancer metastases, benign liver lesions and hepatocellular carcinoma. This semiautomated segmentation method is based on variational interpolation and 3D minimal path-surface segmentation. Total and subsegmental liver volumes were segmented from contrast-enhanced CT images in venous phase. Two image analysts independently performed semiautomated segmentations and two other image analysts performed manual segmentations. Repeatability and agreement of both methods were evaluated with intraclass correlation coefficients (ICC) and Bland-Altman analysis. Interaction time was recorded for both methods. Bland-Altman analysis revealed an intrareader agreement of -1 ± 27 mL (mean ± 1.96 standard deviation) with ICC of 0.999 (P < .001) for manual segmentation and 12 ± 97 mL with ICC of 0.991 (P < .001) for semiautomated segmentation. Bland-Altman analysis revealed an interreader agreement of -4 ± 22 mL with ICC of 0.999 (P < .001) for manual segmentation and 5 ± 98 mL with ICC of 0.991 (P < .001) for semiautomated segmentation. Intermethod agreement was found to be 3 ± 120 mL with ICC of 0.988 (P < .001). Mean interaction time was 34.3 ± 16.7 minutes for the manual method and 8.0 ± 1.2 minutes for the semiautomated method (P < .001). A semiautomated segmentation method can substantially shorten interaction time while preserving a high repeatability and agreement with manual segmentation.

Chronic Statin Administration May Attenuate Early Anthracycline-Associated Declines in Left Ventricular Ejection Function

BackgroundRecent studies have shown an association between statin therapy and a reduced risk of heart failure among breast cancer survivors. Our goal was to evaluate whether statin therapy for prevention of cardiovascular (CV) disease would ameliorate declines in the left ventricular ejection fraction (LVEF) that is often observed during anthracycline-based chemotherapy (Anth-bC). MethodsThere were 51 participants (33 women and 18 men, aged 48 ± 2 years). We obtained cardiovascular magnetic resonance imaging (CMRI) measurements of LVEF before and 6 months after initiation of Anth-bC for patients with breast cancer, leukemia, or lymphoma. Fourteen individuals received statin therapy, and 37 patients received no statins. MR image analysts were blinded to participant identifiers. ResultsIndividuals receiving statins were older and often had diabetes mellitus (DM), hypertension (HTN), and hyperlipidemia (HLD). For those receiving statins, LVEF was 56.6% ± 1.4% at baseline and 54.1% ± 1.3% 6 months after initiating anthracycline treatment (P = 0.15). For those not receiving statins, LVEF was 57.5% ± 1.4% at baseline and decreased to 52.4% ± 1.2% over a similar 6-month interval (P = 0.0003). In a multivariable model accounting for age, sex, DM, HTN, HLD, and cumulative amount of anthracycline received, LVEF remained unchanged in participants receiving a statin (+1.1% ± 2.6%) vs a −6.5% ± 1.5% decline among those not receiving a statin (P = 0.03). ConclusionsThese data highlight the finding that individuals receiving statin therapy for prevention of cardiovascular disease may experience less deterioration in LVEF with early receipt of Anth-bC than individuals not receiving statins. Further studies with large numbers of participants are warranted to determine if statins protect against LVEF decline in patients receiving Anth-bC.

Relative accuracy of spin-image-based registration of partial capitate bones in 4DCT of the wrist

In image-based biomechanical analyses, registration transformations are the data of interest. In dynamic four-dimensional computed tomography (4DCT) imaging, the capitate is often partially imaged. While alignment of incomplete objects poses a significant registration challenge, the established spin-image surface-matching algorithm can be utilised to align two surfaces representing disparate but overlapping portions of an object. For this reason, the spin-image algorithm was chosen for the registration of partial bone geometry in 4DCT of the wrist. Registrations were performed on 11 4DCT datasets using complete and partial capitate meshes generated by cropping complete meshes. Relative accuracy was assessed as the difference between partial- and complete-geometry registrations. Accurate registration of partial capitate geometry was achieved with 55% of the proximal capitate geometry on average, and in some cases as little as 35%. Requisite geometry depends on feature salience and imaging resolution; however, the spin-image algorithm should be considered a valuable tool for biomechanists and image analysts.

Trends in Commercial Radar Remote Sensing Industry [Industrial Profiles

The last two decades have seen an unprecedented development in the satellite-based Earth observation industry. The combination of an increasing number of operational satellites, the higher resolution of the acquired data and the advances in the processing techniques have enabled a wider adoption of satellite data and the development of a diverse range of products and applications. Although the market is still strongly biased toward electro-optically derived imagery a rising tide of acceptance and usage of satellite derived Synthetic Aperture Radar (SAR) data can be noticed over the last few years. This trend is a result of the increasing availability of commercial SAR satellite data, development of sophisticated processing and analysis tools and industry driven training effort conducted to familiarize image analysts with the specifics of SAR imagery, its interpretation and its utility.

Sampling theory and automated simulations for vertical sections, applied to human brain

In recent years, there have been substantial developments in both magnetic resonance imaging techniques and automatic image analysis software. The purpose of this paper is to develop stereological image sampling theory (i.e. unbiased sampling rules) that can be used by image analysts for estimating geometric quantities such as surface area and volume, and to illustrate its implementation. The methods will ideally be applied automatically on segmented, properly sampled 2D images - although convenient manual application is always an option - and they are of wide applicability in many disciplines. In particular, the vertical sections design to estimate surface area is described in detail and applied to estimate the area of the pial surface and of the boundary between cortex and underlying white matter (i.e. subcortical surface area). For completeness, cortical volume and mean cortical thickness are also estimated. The aforementioned surfaces were triangulated in 3D with the aid of FreeSurfer software, which provided accurate surface area measures that served as gold standards. Furthermore, a software was developed to produce digitized trace curves of the triangulated target surfaces automatically from virtual sections. From such traces, a new method (called the 'lambda method') is presented to estimate surface area automatically. In addition, with the new software, intersections could be counted automatically between the relevant surface traces and a cycloid test grid for the classical design. This capability, together with the aforementioned gold standard, enabled us to thoroughly check the performance and the variability of the different estimators by Monte Carlo simulations for studying the human brain. In particular, new methods are offered to split the total error variance into the orientations, sectioning and cycloid components. The latter prediction was hitherto unavailable--one is proposed here and checked by way of simulations on a given set of digitized vertical sections with automatically superimposed cycloid grids of three different sizes. Concrete and detailed recommendations are given to implement the methods.

Acceleration of image analyst training with transcranial direct current stimulation.

Humans today are routinely and increasingly presented with vast quantities of data that challenge their capacity for efficient processing. To restore the balance between man and machine, it is worthwhile to explore new methods for enhancing or accelerating this capacity. This study was designed to investigate the efficacy of transcranial DC stimulation (tDCS) to reduce training time and increase proficiency in spatial recognition using a simulated synthetic aperture radar (SAR) task. Twenty-seven Air Force active duty members volunteered to participate in the study. Each participant was assigned to 1 of 3 stimulation groups and received two, 90-min training sessions on a target search and identification task using SAR imagery followed by a test. The tDCS anode was applied to site F10 according to the 10-20 electroencephalographic electrode convention while the cathode was placed on the contralateral bicep. Group 1 received anodal tDCS at 2 mA for 30 min in the first training session and sham tDCS in the second session. Group 2 received the stimulation conditions in the opposite order. Group 3 did not receive stimulation at all. Results showed that participants receiving training plus tDCS attained visual search accuracies ~25% higher than those provided with sham stimulation or no stimulation. However, a corresponding performance improvement was not found in the first training session for the change detection portion of the task. This indicates that experience with the imagery is important in the tDCS-elicited performance improvements in change detection.

Image Analysis Methods to Monitor Natura 2000 Habitats at Regional Scales – the MS. MONINA State Service Example in Schleswig-Holstein, Germany

Summary: The Natura 2000 network of protected sites is one of the means to address the issue of biodiversity conservation in Europe. Protected under the habitat directive, EU member states have to undertake surveillance of habitats and species of community interest and report every six years on habitat range and distribution, conservation status and the future prospects of the habitats within and outside of protected sites. Remote sensing techniques have been applied successfully to monitor habitat changes relevant for Natura 2000 monitoring using multioral satellite image data, but many challenges remain especially outside protected sites to assess the development of habitats over time. A flexible information layer concept was developed within the FP7 project MS.MONINA to address the complex task of monitoring natural habitats. In this paper the new approach to classify grassland land cover classes in Schleswig-Holstein, Germany, will be presented. Based on ecological parameters experts defined simple description models, which were used by image analysts to extract corresponding image features for four different grassland types. Information layer operators were defined to extract image features for subsequent classifications. (Less)

PROCESSING, CATALOGUING AND DISTRIBUTION OF UAS IMAGES IN NEAR REAL TIME

Abstract. Why are UAS such a hype? UAS make the data capture flexible, fast and easy. For many applications this is more important than a perfect photogrammetric aerial image block. To ensure, that the advantage of a fast data capturing will be valid up to the end of the processing chain, all intermediate steps like data processing and data dissemination to the customer need to be flexible and fast as well. GEOSYSTEMS has established the whole processing workflow as server/client solution. This is the focus of the presentation. Depending on the image acquisition system the image data can be down linked during the flight to the data processing computer or it is stored on a mobile device and hooked up to the data processing computer after the flight campaign. The image project manager reads the data from the device and georeferences the images according to the position data. The meta data is converted into an ISO conform format and subsequently all georeferenced images are catalogued in the raster data management System ERDAS APOLLO. APOLLO provides the data, respectively the images as an OGC-conform services to the customer. Within seconds the UAV-images are ready to use for GIS application, image processing or direct interpretation via web applications – where ever you want. The whole processing chain is built in a generic manner. It can be adapted to a magnitude of applications. The UAV imageries can be processed and catalogued as single ortho imges or as image mosaic. Furthermore, image data of various cameras can be fusioned. By using WPS (web processing services) image enhancement, image analysis workflows like change detection layers can be calculated and provided to the image analysts. The processing of the WPS runs direct on the raster data management server. The image analyst has no data and no software on his local computer. This workflow is proven to be fast, stable and accurate. It is designed to support time critical applications for security demands – the images can be checked and interpreted in near real-time. For sensible areas it gives you the possibility to inform remote decision makers or interpretation experts in order to provide them situations awareness, wherever they are. For monitoring and inspection tasks it speeds up the process of data capture and data interpretation. The fully automated workflow of data pre-processing, data georeferencing, data cataloguing and data dissemination in near real time was developed based on the Intergraph products ERDAS IMAGINE, ERDAS APOLLO and GEOSYSTEMS METAmorph!IT. It is offered as adaptable solution by GEOSYSTEMS GmbH.

Genetic Optimization for Associative Semantic Ranking Models of Satellite Images by Land Cover

Associative methods for content-based image ranking by semantics are attractive due to the similarity of generated models to human models of understanding. Although they tend to return results that are better understood by image analysts, the induction of these models is difficult to build due to factors that affect training complexity, such as coexistence of visual patterns in same images, over-fitting or under-fitting and semantic representation differences among image analysts. This article proposes a methodology to reduce the complexity of ranking satellite images for associative methods. Our approach employs genetic operations to provide faster and more accurate models for ranking by semantic using low level features. The added accuracy is provided by a reduction in the likelihood to reach local minima or to overfit. The experiments show that, using genetic optimization, associative methods perform better or at similar levels as state-of-the-art ensemble methods for ranking. The mean average precision (MAP) of ranking by semantic was improved by 14% over similar associative methods that use other optimization techniques while maintaining smaller size for each semantic model.

A critical analysis of methods and procedures applied to satellite image interpretation by image analysts in the South African National Defence Force.

A critical analysis of methods and procedures applied to satellite image interpretation by image analysts in the South African National Defence Force Inaccurate satellite image interpretation for military purposes may result in inadequate image intelligence and hamper the successful implementation of military operations. The image interpretation methods and procedures applied by South African National Defence Force (SANDF) image analysts raised the question whether image interpretation is correctly and accurately performed by SANDF image analysts. This study uses empirical research in the form of a field experiment to determine whether SANDF military analysts accurately execute military image interpretation.

Collaborative damage mapping for emergency response: the role of Cognitive Systems Engineering

Abstract. Remote sensing is increasingly used to assess disaster damage, traditionally by professional image analysts. A recent alternative is crowdsourcing by volunteers experienced in remote sensing, using internet-based mapping portals. We identify a range of problems in current approaches, including how volunteers can best be instructed for the task, ensuring that instructions are accurately understood and translate into valid results, or how the mapping scheme must be adapted for different map user needs. The volunteers, the mapping organizers, and the map users all perform complex cognitive tasks, yet little is known about the actual information needs of the users. We also identify problematic assumptions about the capabilities of the volunteers, principally related to the ability to perform the mapping, and to understand mapping instructions unambiguously. We propose that any robust scheme for collaborative damage mapping must rely on Cognitive Systems Engineering and its principal method, Cognitive Task Analysis (CTA), to understand the information and decision requirements of the map and image users, and how the volunteers can be optimally instructed and their mapping contributions merged into suitable map products. We recommend an iterative approach involving map users, remote sensing specialists, cognitive systems engineers and instructional designers, as well as experimental psychologists.