Search

Biomedical images are often referenced for clinical decision support (CDS), educational purposes, and research. The task of automatically finding the images in a scientific article that are most useful for the purpose of determining relevance to a clinical situation is traditionally done using text and is quite challenging. We propose to improve this by associating image features from the entire image and from relevant regions of interest with biomedical concepts described in the figure caption or discussion in the article. However, images used in scientific article figures are often composed of multiple panels where each sub-figure (panel) is referenced in the caption using alphanumeric labels, e.g. Figure 1(a), 2(c), etc. It is necessary to separate individual panels from a multi-panel figure as a first step toward automatic annotation of images. In this work we present methods that add make robust our previous efforts reported here. Specifically, we address the limitation in segmenting figures that do not exhibit explicit inter-panel boundaries, e.g. illustrations, graphs, and charts. We present a novel hybrid clustering algorithm based on particle swarm optimization (PSO) with fuzzy logic controller (FLC) to locate related figure components in such images. Results from our evaluation are very promising with 93.64% panel detection accuracy for regular (non-illustration) figure images and 92.1% accuracy for illustration images. A computational complexity analysis also shows that PSO is an optimal approach with relatively low computation time. The accuracy of separating these two type images is 98.11% and is achieved using decision tree.

Abstract: This paper presents some advances in discrimination and fusion algorithms using metal detector (MD) and ground penetrating radar (GPR) sensors in a robotic wand unit. Previously investigated spatially distributed features are extended and fused with discrete wavelet transform representations of MD data. A multilayer perceptron technique is then applied to discriminate between land mine and land mine-like objects based on the wavelet coefficient and time domain features separately. Using MD wavelet and time domain fusion, the probability of false alarms is reduced by 46.0% and 18.0% over the wavelet and time domain models, respectively, at 0.95 probability of detection. Fusion results are presented for the MD and GPR sensors to demonstrate that the two sensors provide complementary information for significantly reducing the probability of false alarm. Blind test results from a government test facility are presented to evaluate the effectiveness of the algorithms.

Biomedical journal articles contain a variety of image types that can be broadly classified into two categories: regular images, and graphical images. Graphical images can be further classified into four classes: diagrams, statistical figures, flow charts, and tables. Automatic figure type identification is an important step toward improved multimodal (text + image) information retrieval and clinical decision support applications. This paper describes a feature-based learning approach to automatically identify these four graphical figure types. We apply Evolutionary Algorithm (EA), Binary Particle Swarm Optimization (BPSO) and a hybrid of EA and BPSO (EABPSO) methods to select an optimal subset of extracted image features that are then classified using a Support Vector Machine (SVM) classifier. Evaluation performed on 1038 figure images extracted from ten BioMedCentral®journals with the features selected by EABPSO yielded classification accuracy as high as 87.5%.

Methods of locating and identifying buried landmines using high-pressure waterjets were investigated. Methods were based on the sound produced when the waterjet strikes a buried object. Three classification techniques were studied, based on temporal, spectral, and a combination of temporal and spectral approaches using weighted density distribution functions, a maximum likelihood approach, and hidden Markov models, respectively. Methods were tested with laboratory data from low-metal content simulants and with field data from inert real landmines. Results show that the sound made when the waterjet hit a buried object could be classified with a 90% detection rate and an 18% false alarm rate. In a blind field test using 3 types of harmless objects and 7 types of landmines, buried objects could be accurately classified as harmful or harmless 60%–90% of the time. High-pressure waterjets may serve as a useful companion to conventional detection and classification methods.