Medical image retrieval with query-dependent feature fusion based on one-class SVM

Due to the huge growth of the World Wide Web, medical images are now available in large numbers in online repositories, and there exists the need to retrieval the images through automatically extracting visual information of the medical images, which is commonly known as content-based image retrieval (CBIR). Since each feature extracted from images just characterizes certain aspect of image content, multiple features are necessarily employed to improve the retrieval performance. Meanwhile, experiments demonstrate that a special feature is not equally important for different image queries. Most of existed feature fusion methods for image retrieval only utilize query independent feature fusion or rely on explicit user weighting. In this paper, we present a novel query dependent feature fusion method for medical image retrieval based on one class support vector machine. Having considered that a special feature is not equally important for different image queries, the proposed query dependent feature fusion method can learn different feature fusion models for different image queries only based on multiply image samples provided by the user, and the learned feature fusion models can reflect the different importances of a special feature for different image queries. The experimental results on the IRMA medical image collection demonstrate that the proposed method can improve the retrieval performance effectively and can outperform existed feature fusion methods for image retrieval.

Image retrieval using noisy query

In conventional content based image retrieval (CBIR) employing relevance feedback, one implicit assumption is that both pure positive and negative examples are available. However it is not always true in the practical applications of CBIR. In this paper, we address a new problem of image retrieval using several unclean positive examples, named noisy query, in which some mislabeled images or weak relevant images present. The proposed image retrieval scheme measures the image similarity by combining multiple feature distances. Incorporating data cleaning and noise tolerant classifier, a twostep strategy is proposed to handle noisy positive examples. Experiments carried out on a subset of Corel image collection show that the proposed scheme outperforms the competing image retrieval schemes.

QDFA : query-dependent feature aggregation for medical image retrieval

We propose a novel query-dependent feature aggregation (QDFA) method for medical image retrieval. The QDFA method can learn an optimal feature aggregation function for a multi-example query, which takes into account multiple features and multiple examples with different importance. The experiments demonstrate that the QDFA method outperforms three other feature aggregation methods.

A new query dependent feature fusion approach for medical image retrieval based on one-class SVM

With the development of the internet, medical images are now available in large numbers in online repositories, and there exists the need to retrieval the medical images in the content-based ways through automatically extracting visual information of the medical images. Since a single feature extracted from images just characterizes certain aspect of image content, multiple features are necessarily employed to improve the retrieval performance. Furthermore, a special feature is not equally important for different image queries since a special feature has different importance in reflecting the content of different images. However, most existed feature fusion methods for image retrieval only utilize query independent feature fusion or rely on explicit user weighting. In this paper, based on multiply query samples provided by the user, we present a novel query dependent feature fusion method for medical image retrieval based on one class support vector machine. The proposed query dependent feature fusion method for medical image retrieval can learn different feature fusion models for different image queries, and the learned feature fusion models can reflect the different importance of a special feature for different image queries. The experimental results on the IRMA medical image collection demonstrate that the proposed method can improve the retrieval performance effectively and can outperform existed feature fusion methods for image retrieval.

Scene extraction in motion pictures

This paper addresses the challenge of bridging the semantic gap between the rich meaning users desire when they query to locate and browse media and the shallowness of media descriptions that can be computed in today's content management systems. To facilitate high-level semantics-based content annotation and interpretation, we tackle the problem of automatic decomposition of motion pictures into meaningful story units, namely scenes. Since a scene is a complicated and subjective concept, we first propose guidelines from fill production to determine when a scene change occurs. We then investigate different rules and conventions followed as part of Fill Grammar that would guide and shape an algorithmic solution for determining a scene. Two different techniques using intershot analysis are proposed as solutions in this paper. In addition, we present different refinement mechanisms, such as film-punctuation detection founded on Film Grammar, to further improve the results. These refinement techniques demonstrate significant improvements in overall performance. Furthermore, we analyze errors in the context of film-production techniques, which offer useful insights into the limitations of our method.

Distributed query processing for mobile surveillance

Addressing core issues in mobile surveillance, we present an architecture for querying and retrieving distributed, semi-permanent multi-modal data through challenged networks with limited connectivity. The system provides a rich set of queries for spatio-temporal querying in a surveillance context, and uses the network availability to provide best quality of service. It incrementally and adaptively refines the query, using data already retrieved that exists on static platforms and on-demand data that it requests from mobile platforms. We demonstrate the system using a real surveillance system on a mobile 20 bus transport network coupled with static bus depot infrastructure. In addition, we show the robustness of the system in handling different conditions in the underlying infrastructure by running simulations on a real, but historic dataset collected in an offline manner.

Combining NL processing and video data to query American football

We explore the use of natural language understanding and image processing to index and query American Football tapes. We present a model for representing spatio-temporal characteristics of multiple objects in dynamic scenes in this domain, and a recognition system which uses the model to recognise American Football plays.

Optimising errors in signaling corporate collapse using MCCCRA

Purpose – The purpose of this paper is to put forward an innovative approach for reducing the variation between Type I and Type II errors in the context of ratio-based modeling of corporate collapse, without compromising the accuracy of the predictive model. Its contribution to the literature lies in resolving the problematic trade-off between predictive accuracy and variations between the two types of errors.

Design/methodology/approach – The methodological approach in this paper – called MCCCRA – utilizes a novel multi-classification matrix based on a combination of correlation and regression analysis, with the former being subject to optimisation criteria. In order to ascertain its accuracy in signaling collapse, MCCCRA is empirically tested against multiple discriminant analysis (MDA).

Findings – Based on a data sample of 899 US publicly listed companies, the empirical results indicate that in addition to a high level of accuracy in signaling collapse, MCCCRA generates lower variability between Type I and Type II errors when compared to MDA.

Originality/value – Although correlation and regression analysis are long-standing statistical tools, the optimisation constraints that are applied to the correlations are unique. Moreover, the multi-classification matrix is a first in signaling collapse. By providing economic insight into more stable financial modeling, these innovations make an original contribution to the literature.

Reported medication errors after introducing an electronic medication management system

Aims and objectives
To explore the effects of introducing an electronic medication management system on reported medication errors.
Background
Computerised medication management systems have been found to improve medication safety; however, introducing medication management system into healthcare environments can create unanticipated or new problems and opportunities for medication error.
Design
Descriptive analysis of medication error reports.
Methods
This was a retrospective analysis of 359 incident reports drawn from the period of 1 May 2005–30 April 2006 across two hospital sites of a single not-for-profit private health service located in metropolitan Melbourne. Site A used a conventional pen and paper system for medication management, and Site B had introduced a computerised medication management system.
Results
Most medication errors occurred at the nurse administration (71·5%) and prescribing (16·4%) stages of delivery. The most common medication error type reported at Site A was omission (33%), and at Site B was wrong documentation (24·2%). A higher proportion of errors at the prescribing phase, and less nurse administration errors, were detected at Site B where the medication management system was in use. The incidence of other, less frequent errors was similar across the two hospital sites.
Conclusions
This examination of medication error reports suggests there are differences in the types of medication errors that are reported in association with the introduction of electronic medication management system compared to pen and paper system systems. The findings provide a new insight into the effects of introducing an electronic medication management system on the types of medication errors reported.
Relevance to clinical practice
The findings provide a new insight into the types of medication errors that are reported during implementation of an electronic medication management system. Extra support for physicians prescribing practices should be considered.

Recruitment into diabetes prevention programs : what is the impact of errors in self-reported measures of obesity?

Background
Error in self-reported measures of obesity has been frequently described, but the effect of self-reported error on recruitment into diabetes prevention programs is not well established. The aim of this study was to examine the effect of using self-reported obesity data from the Finnish diabetes risk score (FINDRISC) on recruitment into the Greater Green Triangle Diabetes Prevention Project (GGT DPP).

Methods
The GGT DPP was a structured group-based lifestyle modification program delivered in primary health care settings in South-Eastern Australia. Between 2004–05, 850 FINDRISC forms were collected during recruitment for the GGT DPP. Eligible individuals, at moderate to high risk of developing diabetes, were invited to undertake baseline tests, including anthropometric measurements performed by specially trained nurses. In addition to errors in calculating total risk scores, accuracy of self-reported data (height, weight, waist circumference (WC) and Body Mass Index (BMI)) from FINDRISCs was compared with baseline data, with impact on participation eligibility presented.

Results
Overall, calculation errors impacted on eligibility in 18 cases (2.1%). Of n = 279 GGT DPP participants with measured data, errors (total score calculation, BMI or WC) in self-report were found in n = 90 (32.3%). These errors were equally likely to result in under- or over-reported risk. Under-reporting was more common in those reporting lower risk scores (Spearman-rho = −0.226, p-value < 0.001). However, underestimation resulted in only 6% of individuals at high risk of diabetes being incorrectly categorised as moderate or low risk of diabetes.

Conclusions
Overall FINDRISC was found to be an effective tool to screen and recruit participants at moderate to high risk of diabetes, accurately categorising levels of overweight and obesity using self-report data. The results could be generalisable to other diabetes prevention programs using screening tools which include self-reported levels of obesity.

Errors in the identification of question types in investigative interviews of children

This study examined the incidence and nature of the errors made by trainee coders during their coding of question types in interviews in which children disclosed abuse. Three groups of trainees (online, postgraduate and police) studied the coding manual before practising their question coding. After this practice, participants were given two-page field transcripts to code in which children disclosed abuse. Their coding was assessed for accuracy; any errors were analysed thematically. The overall error rate was low, and police participants made the fewest errors. Analysis of the errors revealed four common misunderstandings: (1) the use of a ‘wh’ question always denotes a specific cued-recall question; (2) ‘Tell me’ always constitutes an open-ended question; (3) open-ended questions cannot include specific detail; and (4) specific questions cannot elicit elaborate responses. An analysis of coding accuracy in the one group who were able to practise question coding over time revealed that practice was essential for trainees to maintain their accuracy. Those who did not practise decreased in coding accuracy. This research shows that trainees need more than a coding manual; they must demonstrate their understanding of question codes through practice training tasks. Misunderstandings about questions need to be elicited and corrected so that accurate codes are used in future tasks.