SEARCH · PubMed Health

Results for “generative artificial intelligence”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,099 records · Page 61Linked to original sources

PubMed2004

X-ray CT high-density artefact suppression in the presence of bones.

This paper presents a novel method of reducing x-ray CT high-density artefacts generated by metal objects when abundant bone structures are present in the region of interest. This method has an advantage over previously proposed methods since it heavily suppresses the metal artefacts without introducing extra bone artefacts. The method of suppression requires that bone pixels are isolated and segmented by thresholding. Then artificial CT numbers are assigned to the bone pixels so that their projection profiles are smooth and thus can be properly simulated by a polynomial interpolation. The projection profile of the metal object is then removed to fully suppress the artefacts. The resulting processed profile is fed to a reconstruction routine and the previously preserved bone pixels added back. The new method utilizes two important features of the CT image with metal artefacts: (a) metal and bone pixels are not severely affected by the high-density artefacts and (b) the high-density artefacts can be located in specific projection channels in the profile domain, although they are spread out in the image domain. This suppression method solves the problem of CT image artefacts arising from metal objects in the body. It has the potential to greatly improve diagnostic CT imaging in the presence of these objects and treatment planning that utilizes CT for patients with metal applicators (e.g., brachytherapy for cervix cancer and prostate cryotherapy).

Algorithms↗

PubMed2005

Dynamic Trees for unsupervised segmentation and matching of image regions.

We present a probabilistic framework--namely, multiscale generative models known as Dynamic Trees (DT)--for unsupervised image segmentation and subsequent matching of segmented regions in a given set of images. Beyond these novel applications of DTs, we propose important additions for this modeling paradigm. First, we introduce a novel DT architecture, where multilayered observable data are incorporated at all scales of the model. Second, we derive a novel probabilistic inference algorithm for DTs--Structured Variational Approximation (SVA)--which explicitly accounts for the statistical dependence of node positions and model structure in the approximate posterior distribution, thereby relaxing poorly justified independence assumptions in previous work. Finally, we propose a similarity measure for matching dynamic-tree models, representing segmented image regions, across images. Our results for several data sets show that DTs are capable of capturing important component-subcomponent relationships among objects and their parts, and that DTs perform well in segmenting images into plausible pixel clusters. We demonstrate the significantly improved properties of the SVA algorithm--both in terms of substantially faster convergence rates and larger approximate posteriors for the inferred models--when compared with competing inference algorithms. Furthermore, results on unsupervised object recognition demonstrate the viability of the proposed similarity measure for matching dynamic-structure statistical models.

Algorithms↗

PubMed1998

Knowledge-based support for the participatory design and implementation of shift systems.

OBJECTIVES: This study developed a knowledge-based software system to support the participatory design and implementation of shift systems as a joint planning process including shift workers, the workers' committee, and management. METHODS: The system was developed using a model-based approach. During the 1st phase, group discussions were repeatedly conducted with 2 experts. Thereafter a structure model of the process was generated and subsequently refined by the experts in additional semistructured interviews. Next, a factual knowledge base of 1713 relevant studies was collected on the effects of shift work. Finally, a prototype of the knowledge-based system was tested on 12 case studies. RESULTS: During the first 2 phases of the system, important basic information about the tasks to be carried out is provided for the user. During the 3rd phase this approach uses the problem-solving method of case-based reasoning to determine a shift rota which has already proved successful in other applications. It can then be modified in the 4th phase according to the shift workers' preferences. The last 2 phases support the final testing and evaluation of the system. The application of this system has shown that it is possible to obtain shift rotas suitable to actual problems and representative of good ergonomic solutions. CONCLUSIONS: A knowledge-based approach seems to provide valuable support for the complex task of designing and implementing a new shift system. The separation of the task into several phases, the provision of information at all stages, and the integration of all parties concerned seem to be essential factors for the success of the application.

Artificial Intelligence↗

PubMed2004

An architecture for biological information extraction and representation.

MOTIVATIONS: Technological advances in biomedical research are generating a plethora of heterogeneous data at a high rate. There is a critical need for extraction, integration and management tools for information discovery and synthesis from these heterogeneous data. RESULTS: In this paper, we present a general architecture, called ALFA, for information extraction and representation from diverse biological data. The ALFA architecture consists of: (i) a networked, hierarchical, hyper-graph object model for representing information from heterogeneous data sources in a standardized, structured format; and (ii) a suite of integrated, interactive software tools for information extraction and representation from diverse biological data sources. As part of our research efforts to explore this space, we have currently prototyped the ALFA object model and a set of interactive software tools for searching, filtering, and extracting information from scientific text. In particular, we describe BioFerret, a meta-search tool for searching and filtering relevant information from the web, and ALFA Text Viewer, an interactive tool for user-guided extraction, disambiguation, and representation of information from scientific text. We further demonstrate the potential of our tools in integrating the extracted information with experimental data and diagrammatic biological models via the common underlying ALFA representation. CONTACT: aditya_vailaya@agilent.com.

Abstracting and Indexing↗

PubMed2005

Java Web Start based software for automated quantitative nuclear analysis of prostate cancer and benign prostate hyperplasia.

BACKGROUND: Androgen acts via androgen receptor (AR) and accurate measurement of the levels of AR protein expression is critical for prostate research. The expression of AR in paired specimens of benign prostate and prostate cancer from 20 African and 20 Caucasian Americans was compared to demonstrate an application of this system. METHODS: A set of 200 immunopositive and 200 immunonegative nuclei were collected from the images using a macro developed in Image Pro Plus. Linear Discriminant and Logistic Regression analyses were performed on the data to generate classification coefficients. Classification coefficients render the automated image analysis software independent of the type of immunostaining or image acquisition system used. The image analysis software performs local segmentation and uses nuclear shape and size to detect prostatic epithelial nuclei. AR expression is described by (a) percentage of immunopositive nuclei; (b) percentage of immunopositive nuclear area; and (c) intensity of AR expression among immunopositive nuclei or areas. RESULTS: The percent positive nuclei and percent nuclear area were similar by race in both benign prostate hyperplasia and prostate cancer. In prostate cancer epithelial nuclei, African Americans exhibited 38% higher levels of AR immunostaining than Caucasian Americans (two sided Student's t-tests; P < 0.05). Intensity of AR immunostaining was similar between races in benign prostate. CONCLUSION: The differences measured in the intensity of AR expression in prostate cancer were consistent with previous studies. Classification coefficients are required due to non-standardized immunostaining and image collection methods across medical institutions and research laboratories and helps customize the software for the specimen under study. The availability of a free, automated system creates new opportunities for testing, evaluation and use of this image analysis system by many research groups who study nuclear protein expression.

Algorithms↗

PubMed2004

Estimating degradation model parameters using neighborhood pattern distributions: an optimization approach.

Noise models are crucial for designing image restoration algorithms, generating synthetic training data, and predicting algorithm performance. There are two related but distinct estimation scenarios. The first is model calibration, where it is assumed that the input ideal bitmap and the output of the degradation process are both known. The second is the general estimation problem, where only the image from the output of the degradation process is given. While researchers have addressed the problem of calibration of models, issues with the general estimation problems have not been addressed in the literature. In this paper, we describe a parameter estimation algorithm for a morphological, binary, page-level image degradation model. The inputs to the estimation algorithm are 1) the degraded image and 2) information regarding the font type (italic, bold, serif, sans serif). We simulate degraded images using our model and search for the optimal parameter by looking for a parameter value for which the local neighborhood pattern distributions in the simulated image and the given degraded image are most similar. The parameter space is searched using a direct search optimization algorithm. We use the p-value of the Kolmogorov-Smirnov test as the measure of similarity between the two neighborhood pattern distributions. We show results of our algorithm on degraded document images.

Algorithms↗

PubMedDate not supplied

Evaluation of a new method for cardiovascular reasoning.

OBJECTIVE: Evaluate the accuracy of the detailed diagnostic reasoning of the Heart Failure Program incorporating a new mechanism to handle temporal relationships and severity constraints. DESIGN: Tools were developed to summarize diagnoses and automatically generate evaluation forms. Five expert cardiologists were asked to review the reasoning of the program, with two analyzing each case. Cases were gathered retrospectively for diversity and difficulty and 26 randomly selected cases were evaluated. The underlying issues were identified and classified. RESULTS: Both reviewers rated the first diagnosis correct in 25% of the cases and at least one rated it wrong in 10%. Analyzing the detailed reasoning, 137 issues were raised, about 5.3 per case. Of these, 53% were possible concerns raised by one reviewer. Of the 5.3 issues per case, 2.5 were attributable to controversies, misunderstandings, or mistakes; 1 was due to the overly simplistic representation of the summaries; and 1.8 were issues related to the program. CONCLUSION: Overall, the program is capable of providing high-quality detailed diagnostic hypotheses for complex cardiovascular cases. The results highlight several issues: 1) the difficulty of effectively summarizing hypotheses, 2) the nature of a physician's causal explanation, and 3) some problems in evaluating detailed diagnostic reasoning. The mistakes the program made imply that some additional refinement is needed but that the reasoning mechanisms developed can support the appropriate reasoning. The appropriate next step is a prospective evaluation addressing the program's usefulness.

Aged↗

PubMed1995

Prudent expert systems with credentials: managing the expertise of decision support systems.

'Black box' expert systems (ES) are mistrusted by clinicians. Errors generated by medical ES are also a significant cause for concern. We report new ES properties--prudence and credentials--that improve error management and underpin a new approach for improving the credibility of ES for clinical users. Prudent ES modify their output according to past experience. For a knowledge base built from 1610 cases, feature exception prudence (FEP) detected all interpretation errors (100% sensitivity for error detection). Although the false positive rate for FEP was high (47%), the 100% sensitivity meant that the 53% of cases that did not produce flags could be exempted from human validation. As more cases are processed, fewer cases should need human validation. Feature recognition prudence (FRP), a property of ripple down rules (RDR), proposed the correct alternative conclusion in 14% of incorrectly interpreted cases. Human expert validation of the flagged cases enabled context-sensitive credentials (accuracy, incidence and specificity of a given conclusion) to accumulate. Credentials should enable the user to judge the credibility of the ES output. An error management strategy based on credentialed, prudent ES should reduce the impact of error in the clinical environment. The empowerment of clinicians to critically evaluate ES credibility may facilitate greater confidence in, and acceptance of, ES by clinicians.

Artificial Intelligence↗

PubMed2006

Self-calibrating 3D-ultrasound-based bone registration for minimally invasive orthopedic surgery.

Intraoperative freehand three-dimensional (3-D) ultrasound (3D-US) has been proposed as a noninvasive method for registering bones to a preoperative computed tomography image or computer-generated bone model during computer-aided orthopedic surgery (CAOS). In this technique, an US probe is tracked by a 3-D position sensor and acts as a percutaneous device for localizing the bone surface. However, variations in the acoustic properties of soft tissue, such as the average speed of sound, can introduce significant errors in the bone depth estimated from US images, which limits registration accuracy. We describe a new self-calibrating approach to US-based bone registration that addresses this problem, and demonstrate its application within a standard registration scheme. Using realistic US image data acquired from 6 femurs and 3 pelves of intact human cadavers, and accurate Gold Standard registration transformations calculated using bone-implanted fiducial markers, we show that self-calibrating registration is significantly more accurate than a standard method, yielding an average root mean squared target registration error of 1.6 mm. We conclude that self-calibrating registration results in significant improvements in registration accuracy for CAOS applications over conventional approaches where calibration parameters of the 3D-US system remain fixed to values determined using a preoperative phantom-based calibration.

Algorithms↗

PubMedDate not supplied

The use of machine learning program LERS-LB 2.5 in knowledge acquisition for expert system development in nursing.

LERS-LB (Learning from Examples using Rough Sets Lower Boundaries) is a computer program based on rough set theory for knowledge acquisition, which extracts patterns from real-world data in generating production rules for expert system development. From LERS-LB evaluation of an SPSS-X data file containing data for recovery room patients, it was concluded that both statistical data files and existing databases can be converted to decision-table format needed by LERS-LB, but it is less desirable to work with statistical files than a well-developed database. It was also concluded that choosing a well-developed database and checking it thoroughly for accuracy and completeness should be done before running LERS-LB, or other learning programs, to avoid problems with data errors. Using rough set theory and a technique called 'dropping conditions' LERS-LB offers, at least in theory, a possible method for identifying which data items are critical to nursing practice. Further research and continued LERS-LB program enhancements still may help with identifying critical data items versus redundant data for nursing practice. LERS-LB, and other learning programs, offer techniques which will help reduce the knowledge acquisition bottleneck in nursing expert system development. It is doubtful, however, that learning programs will eliminate the need for involving domain experts in evaluating rules and expert systems for clinical decision support.

Artificial Intelligence↗

PubMed1989

Shape the implant to the patient. A rationale for the use of custom-fit cementless total hip implants.

Most authors agree that "fit" and "fill" of the endosteal canal is critical to the success of cementless femoral components. Yet, there exists a wide variation to the anatomic size and shape of this canal in the normal population. This paper reviews the rationale and the efficacy of using a computed tomography-generated CAD/CAM custom femoral component in cementless total hip arthroplasty. One hundred fifty-six cases (81 primary and 75 revisions) were reviewed with follow-up time of six weeks to three years (mean, 22 months). A subset of 48 hips (25 primary and 23 revisions) has been followed a minimum of two years. For the primary hips, the custom group was found to have statistically higher Harris pain scores (less pain) at all follow-up intervals as compared to a prior series by the same surgeon using an off-the-shelf (OTS) prosthesis. Revision customized hip implants had lower Harris pain and total scores than primary custom hips, but 80% were in the none or slight pain category. In revision cases, the use of custom components decreased the need for structural bone grafting and achieved stability on host bone in situations in which it was not possible using OTS components. Complications included failure by aseptic loosening of one primary and one revision case. Initial subsidence of more than 3 mm of the collarless custom design occurred in 8%, the majority being in revision cases. All cases but one appear to have stabilized. Subsidence occurred mainly in cases done early in the series.(ABSTRACT TRUNCATED AT 250 WORDS)

Artificial Intelligence↗

PubMed1988

Theory formation in postulating enzyme kinetic mechanisms: reasoning with constraints.

This paper reports on a prototype system for modeling and analyzing the expert reasoning involved in postulating enzyme kinetic mechanisms. It involves data-driven, theory-driven, and analogical components of reasoning within a generate-and-test cycle. Its central component is a set of domain-specific "filters" for matching experimentally and theoretically derived constraints. The input to the system consists of an abstracted qualitative description of an experiment and prior knowledge reported in the literature. Its output shows how the results match those expected for a set of postulated reaction mechanism models and also provides a trace of which features do or do not match each of the candidate topological models. Results, constraints, and models are all analyzed and compared to those from other, similar experiments. We deduced rules for interpreting the qualitative features of enzyme kinetic experiments from natural language descriptions in the literature and verified that the rules were correct by predicting the results for typical mechanisms. We obtained the correct behavior for all 37 states of a complex enzyme mechanism involving three substrates and three products. We tested our system on data from several published reports dealing with the enzyme hexokinase and obtained detailed listings of the differences in conclusions and interpretation reported in several journal articles. This system, which provides qualitative representations of enzyme kinetic results, should facilitate further experimentation on theory formation in enzyme kinetics and lead to more efficient experimental designs.

Artificial Intelligence↗

PubMed2006

Stock trading using RSPOP: a novel rough set-based neuro-fuzzy approach.

This paper investigates the method of forecasting stock price difference on artificially generated price series data using neuro-fuzzy systems and neural networks. As trading profits is more important to an investor than statistical performance, this paper proposes a novel rough set-based neuro-fuzzy stock trading decision model called stock trading using rough set-based pseudo outer-product (RSPOP) which synergizes the price difference forecast method with a forecast bottleneck free trading decision model. The proposed stock trading with forecast model uses the pseudo outer-product based fuzzy neural network using the compositional rule of inference [POPFNN-CRI(S)] with fuzzy rules identified using the RSPOP algorithm as the underlying predictor model and simple moving average trading rules in the stock trading decision model. Experimental results using the proposed stock trading with RSPOP forecast model on real world stock market data are presented. Trading profits in terms of portfolio end values obtained are benchmarked against stock trading with dynamic evolving neural-fuzzy inference system (DENFIS) forecast model, the stock trading without forecast model and the stock trading with ideal forecast model. Experimental results showed that the proposed model identified rules with greater interpretability and yielded significantly higher profits than the stock trading with DENFIS forecast model and the stock trading without forecast model.

Algorithms↗

PubMed2005

Predicting clinical variable from MRI features: application to MMSE in MCI.

The ability to predict a clinical variable from automated analysis of single, cross-sectional T1-weighted (T1w) MR scans stands to improve the management of patients with neurological diseases. We present a methodology for predicting yearly Mini-Mental Score Examination (MMSE) changes in Mild Cognitive Impairment (MCI) patients. We begin by generating a non-pathological, multidimensional reference space from a group of 152 healthy volunteers by Principal Component Analyses of (i) T1w MR intensity of linearly registered Volumes of Interest (VOI); and (ii) trace of the deformation fields of nonlinearly registered VOIs. We use multiple regression to build linear models from eigenvectors where the projection eigencoordinates of patient data in the reference space are highly correlated with the clinical variable of interest. In our cohort of 47 MCI patients, composed of 16 decliners, 26 stable and 5 improvers (based on MMSE at 1 yr follow-up), there was a significant difference (P = 0.0003) for baseline MMSE scores between decliners and improvers, but no other differences based on age or sex. First, we classified our three groups using leave-one-out, forward stepwise linear discriminant analyses of the projection eigencoordinates with 100% accuracy. Next, we compared various linear models by computing F-statistics on the residuals of predicted vs actual values. The best model was based on 10 eigenvectors + baseline MMSE, with predicted yearly changes highly correlated (r = 0.6955) with actual data. Prospective study of an independent cohort of patients is the next logical step towards establishing this promising technique for clinical use.

Artificial Intelligence↗

PubMed2006

Association of different prediction methods for determination of the efficiency and selectivity on neuron-based sensors.

A technique has been developed to determine the efficiency and the selectivity of a single neuron-based sensor in identifying the nature of the chemical agents in an unknown sample. This has been achieved by exploiting the unique electrical identifiers, also known as "signature patterns", generated by the neuronal cell membrane. These were generated based on the variations to the extracellular electrical activity, due to the effect of a broad range of chemical agents. We demonstrate the prediction capability of the sensor in identifying the nature of an unknown test sample from a combination of three chemical agents, namely, ethanol, pyrethroid, and hydrogen peroxide. This was achieved through a two-step process. The first step was experimentally achieved by in situ recording of the changes to the extracellular electrical activity from the sensing sites or the array of microelectrodes that form the platform for patterning neurons. Simultaneous optical characterization of the cell array during the sensing process was performed to identify the associated physiological changes. The second step was mathematical and was based on developing a library of signature patterns for a set of concentrations of the various combinations of the three chemical agents. Two variants of the nearest neighbor algorithm scheme - (a) partial distance search method, and (b) search tree method, were implemented for the accurate detection of all the components with varying concentrations in the test samples of unknown nature. This technique exhibits reliability in identification up to parts-per-billion (ppb) sensitivity. The capability of standardization of this technique for potential commercial applications is also discussed.

Action Potentials↗

PubMed2004

[Voice recognition applied to ICU medical reports].

OBJECTIVE: Medical reports in ICU are usually tape recorded and then typed by transcriptionist. Digital voice recognition (DVR) has considerably improved and may be a time saving method. The purpose of this study was to test IBM Viavoice (Pro Edition, Release 8, French version) for the generation of medical reports in a surgical intensive care unit. METHODS: A senior resident (Int) in anaesthesia, a senior academic (AS) anaesthesiologist and a confirmed medical secretary (Sec) all working in the same surgical ICU participated to the study. First, Int and AS trained Viavoice according to the software requirements. Then 60 medical records (text files) served to automatically train the software. For the study, we prospectively used 20 consecutive medical records, which had been dictated on tape by either Int or AS. The transcription time by Sec, the number of typing errors and time needed for corrections were recorded. Int and AS with Viavoice processed the same records. Turnaround time and errors were assessed for comparison to Sec work. RESULTS: Accuracy was 97, 92 and 95% for Sec, Int and AS, respectively. Transcription by Viavoice was always significantly faster whatever the speaker than conventional transcription by a secretary. However, Int required significantly longer time to correct recorded text than did Sec and AS. The total turnaround time was significantly faster for AS than Sec (p < 0.01) but Int did not differ from Sec. The performance of Viavoice did not rise with the number of records. CONCLUSION: Viavoice is as efficient as a confirmed transcriptionist.

Anesthesia↗

PubMed2006

Selection and parameterization of cortical neurons for neuroprosthetic control.

When designing neuroprosthetic interfaces for motor function, it is crucial to have a system that can extract reliable information from available neural signals and produce an output suitable for real life applications. Systems designed to date have relied on establishing a relationship between neural discharge patterns in motor cortical areas and limb movement, an approach not suitable for patients who require such implants but who are unable to provide proper motor behavior to initially tune the system. We describe here a method that allows rapid tuning of a population vector-based system for neural control without arm movements. We trained highly motivated primates to observe a 3D center-out task as the computer played it very slowly. Based on only 10-12 s of neuronal activity observed in M1 and PMd, we generated an initial mapping between neural activity and device motion that the animal could successfully use for neuroprosthetic control. Subsequent tunings of the parameters led to improvements in control, but the initial selection of neurons and estimated preferred direction for those cells remained stable throughout the remainder of the day. Using this system, we have observed that the contribution of individual neurons to the overall control of the system is very heterogeneous. We thus derived a novel measure of unit quality and an indexing scheme that allowed us to rate each neuron's contribution to the overall control. In offline tests, we found that fewer than half of the units made positive contributions to the performance. We tested this experimentally by having the animals control the neuroprosthetic system using only the 20 best neurons. We found that performance in this case was better than when the entire set of available neurons was used. Based on these results, we believe that, with careful task design, it is feasible to parameterize control systems without any overt behaviors and that subsequent control system design will be enhanced with cautious unit selection. These improvements can lead to systems demanding lower bandwidth and computational power, and will pave the way for more feasible clinical systems.

Action Potentials↗

PubMedDate not supplied

Ad hoc classification of radiology reports.

OBJECTIVE: The task of ad hoc classification is to automatically place a large number of text documents into nonstandard categories that are determined by a user. The authors examine the use of statistical information retrieval techniques for ad hoc classification of dictated mammography reports. DESIGN: The authors' approach is the automated generation of a classification algorithm based on positive and negative evidence that is extracted from relevance-judged documents. Test documents are sorted into three conceptual bins: membership in a user-defined class, exclusion from the user-defined class, and uncertain. Documentation of absent findings through the use of negation and conjunction, a hallmark of interpretive test results, is managed by expansion and tokenization of these phrases. MEASUREMENTS: Classifier performance is evaluated using a single measure, the F measure, which provides a weighted combination of recall and precision of document sorting into true positive and true negative bins. RESULTS: Single terms are the most effective text feature in the classification profile, with some improvement provided by the addition of pairs of unordered terms to the profile. Excessive iterations of automated classifier enhancement degrade performance because of overtraining. Performance is best when the proportions of relevant and irrelevant documents in the training collection are close to equal. Special handling of negation phrases improves performance when the number of terms in the classification profile is limited. CONCLUSIONS: The ad hoc classifier system is a promising approach for the classification of large collections of medical documents. NegExpander can distinguish positive evidence from negative evidence when the negative evidence plays an important role in the classification.

Algorithms↗