Dermatologists often diagnose or rule out early melanoma by evaluating the follow-up dermoscopic images of skin lesions. However, existing algorithms for early melanoma diagnosis are developed using single time-point images of lesions. Ignoring the temporal, morphological changes of lesions can lead to misdiagnosis in borderline cases. In this study, we propose a framework for automated early melanoma diagnosis using sequential dermoscopic images. To this end, we construct our method in three steps. First, we align sequential dermoscopic images of skin lesions using estimated Euclidean transformations, extract the lesion growth region by computing image differences among the consecutive images, and then propose a spatio-temporal network to capture the dermoscopic changes from aligned lesion images and the corresponding difference images. Finally, we develop an early diagnosis module to compute probability scores of malignancy for lesion images over time. We collected 179 serial dermoscopic imaging data from 122 patients to verify our method. Extensive experiments show that the proposed model outperforms other commonly used sequence models. We also compared the diagnostic results of our model with those of seven experienced dermatologists and five registrars. Our model achieved higher diagnostic accuracy than clinicians (63.69% vs. 54.33%, respectively) and provided an earlier diagnosis of melanoma (60.7% vs. 32.7% of melanoma correctly diagnosed on the first follow-up images). These results demonstrate that our model can be used to identify melanocytic lesions that are at high-risk of malignant transformation earlier in the disease process and thereby redefine what is possible in the early detection of melanoma.

We describe and validate a system for monitoring social contagions on Twitter: social movements, rumors, and emotional outbursts that spread from person to person in a viral manner. We use Twitter streams to monitor the spread of these phenomena through human social and information networks. This system, the contagion monitor, parses Twitter posts to identify emerging phenomena, as captured in hashtags, URLs, words and phrases, or account-handles, and then determines the extent to which a particular phenomenon spreads via the social network (in contrast to its spread via news broadcasts or independent adoption) and locates the contagion within Twitter communities. The monitor approximates the adoption threshold of a social contagion by measuring the fraction of Twitter users who were “infected” by the contagion (e.g., joined a particular social movement) after more than one of their friends had done so. Finally, the monitor makes a judgment about whether the phenomenon has reached critical mass, which is defined as the point where a social contagion begins spreading rapidly and breaches the social boundaries of its early adopter group. We test our prototype monitor on two data sources - an ongoing stream of tweets grouped by user-added hashtags and a collection of posts by a monitored set of Nigerian Twitter users - before productionalizing. We use the Amazon Mechanical Turk platform to evaluate the performance on both data sources. In both cases, we find that our approach successfully distinguishes between high-threshold and low-threshold social contagions.

After playing a five to seven minute character guessing game with a Nao robot, children answered questions about their perceptions of the robot's abilities. Responses from interactions with 30 children, ages eight to twelve, showed that when the robot made an attempt at guessing the participant's character, rather than being stumped and unable to guess, the robot was more likely to be perceived as being able to understand the participant's feelings and able to provide advice. Regardless of their game experience, boys were more likely than girls to feel they could have discussions with the robot about things they could not talk to other people about. This article provides details associated with the implementation of a game used to guess a character the children selected; a twelve question verbally-administered survey that examined their perceptions of the robot; quantitative and qualitative results from the study; and a discussion of the implications, limitations, and future directions of this research.

Entropy, alpha, and anisotropy (H/alpha/A) of the polarimetric target decomposition have been an effective and popular tool for polarimetric synthetic aperture radar (SAR) image analysis and for a geophysical parameter estimation. However, multilook processing can severely affect the values of these parameters. In this paper, a Monte Carlo simulation is used to evaluate and remove the bias generated by the multilook effect on these parameters for various media composed of grassland, forest, and urban returns. Due to insufficient averaging, entropy is underestimated, and anisotropy is overestimated. We also found that the bias in the alpha angle can be either underestimated or overestimated depending on scattering mechanisms. Based on simulation results, efficient bias removal procedures have been developed. In particular, the entropy bias can be precisely corrected, and the amount of correction is independent of the radar frequency and SAR systems. Data from L-band Advanced Land Observing Satellite/phased array type L-band SAR, German Aerospace Research Center (DLR)/enhanced SAR, Jet Propulsion Laboratory (JPL)/airborne SAR, and X-band polarimetric and interferometric SAR are used for demonstration in this paper.

In this paper we investigate the nonstationary behavior of individual polarimetric parameters, e.g. entropy, anisotropy, alpha angle, orientation angle, helicity, etc. We distinguish between parameters that depend solely on the eigenvalues of the standard Cloude-Pottier polarimetric decomposition (span, entropy and anisotropy) and the others that depend on the scattering mechanisms, i.e. the Cloude-Pottier eigenvectors. After producing a series of azimuth subaperture polarimetric images, we apply the polarimetric decomposition to each subaperture frame and identify subaperture frames that show nonstationary scattering. Comparison of several specific targets, both discrete and distributed, will highlight aspects of polarimetric variability with respect to the SAR view angle. We illustrate our results using EMISAR L-band polarimetric SAR data.

Entropy, alpha and anisotropy (H/α/A) of the polarimetric target decomposition has been an effective and popular tool for polarimetric SAR image analysis and geophysical parameter estimation. However, multi-look processing can severely affects the values of these parameters. In this paper, we evaluate the bias problem in H/a/A due to insufficient averaging. We found that the estimated bias is radar frequency dependent. A procedure for bias compensation is proposed. Data from L-band DLR/E-SAR and L-band JPL/AIRSAR, and X-band PI-SAR data are used for demonstration in this study.

Hospital Accident and Emergency (A&E) departments in England have a 4 hour target to treat 98% of patients from arrival to discharge, admission or transfer. Managing resources to meet the target and deliver care across the range of A&E services is a huge challenge for A&E managers. This paper develops an intelligent patient management tool to help managers and clinicians better understand patient length of stay and resources within an A&E area. The developed discrete-event simulation model gives a highlevel representation of ambulance arrivals into A&E. The model facilitates analysis in the following ways: visually interactive software showing patient length of stay in the A&E area; patient activity broken down into sub-groups so that intelligence might be gathered on how sub-groups affect the overall length of stay; understanding the number of patient treatment places and nurse resources required. To support ease of inputs for scenario and sensitivity testing, data is entered into the simulation model (Simul8) via Excel spreadsheets. The model discussed in this paper used patient length of stay grouped by A&E diagnosis codes and was limited to ambulance arrivals. The analysis was derived from A&E attendance in 2004 from an English hospital.

The paper describes three case studies in the lightweight application of formal methods to requirements modeling for spacecraft fault protection systems. The case studies differ from previously reported applications of formal methods in that formal methods were applied very early in the requirements engineering process to validate the evolving requirements. The results were fed back into the projects to improve the informal specifications. For each case study, we describe what methods were applied, how they were applied, how much effort was involved, and what the findings were. In all three cases, formal methods enhanced the existing verification and validation processes by testing key properties of the evolving requirements and helping to identify weaknesses. We conclude that the benefits gained from early modeling of unstable requirements more than outweigh the effort needed to maintain multiple representations.