The adoption of high-accuracy speech recognition algorithms without an effective evaluation of their impact on the target computational resource is impractical for mobile and embedded systems. In this paper, techniques are adopted to minimise the required computational resource for an effective mobile-based speech recognition system. A Dynamic MultiLayer Perceptron speech recognition technique, capable of running in real time on a state-of-the-art mobile device, has been introduced. Even though a conventional hidden Markov model when applied to the same dataset slightly outperformed our approach, its processing time is much higher. The Dynamic Multi-layer Perceptron presented here has an accuracy level of 96.94% and runs significantly faster than similar techniques.

BackgroundOestrogen receptor (ER) positive (luminal) tumours account for the largest proportion of females with breast cancer. Theirs is a heterogeneous disease presenting clinical challenges in managing their treatment. Three main biological luminal groups have been identified but clinically these can be distilled into two prognostic groups in which Luminal A are accorded good prognosis and Luminal B correlate with poor prognosis. Further biomarkers are needed to attain classification consensus. Machine learning approaches like Artificial Neural Networks (ANNs) have been used for classification and identification of biomarkers in breast cancer using high throughput data. In this study, we have used an artificial neural network (ANN) approach to identify DACH1 as a candidate luminal marker and its role in predicting clinical outcome in breast cancer is assessed.Materials and methodsA reiterative ANN approach incorporating a network inferencing algorithm was used to identify ER-associated biomarkers in a publically available cDNA microarray dataset. DACH1 was identified in having a strong influence on ER associated markers and a positive association with ER. Its clinical relevance in predicting breast cancer specific survival was investigated by statistically assessing protein expression levels after immunohistochemistry in a series of unselected breast cancers, formatted as a tissue microarray.ResultsStrong nuclear DACH1 staining is more prevalent in tubular and lobular breast cancer. Its expression correlated with ER-alpha positive tumours expressing PgR, epithelial cytokeratins (CK)18/19 and ‘luminal-like’ markers of good prognosis including FOXA1 and RERG (p<0.05). DACH1 is increased in patients showing longer cancer specific survival and disease free interval and reduced metastasis formation (p<0.001). Nuclear DACH1 showed a negative association with markers of aggressive growth and poor prognosis.ConclusionNuclear DACH1 expression appears to be a Luminal A biomarker predictive of good prognosis, but is not independent of clinical stage, tumour size, NPI status or systemic therapy.

No abstract

Colour information in form analysis is currently under utilised. As technology has advanced and computing costs have reduced, the processing of forms in colour has now become practicable. This paper describes a novel colour-based approach to the extraction of filled data from colour form images.Images are first quantised to reduce the colour complexity and data is extracted by examining the colour characteristics of the images. The improved performance of the proposed method has been verified by comparing the processing time, recognition rate, extraction precision and recall rate to that of an equivalent black and white system.

No abstract

This paper describes how the fault tolerance of the backpropagation algorithm can be used to accommodate the realistic (nonideal) transfer characteristics of the optical communication links used, between neural layers, in optoelectronic neural networks. In particular the authors demonstrate that networks, utilizing MSM (metal-semiconductor-metal) photodiodes (PDs) and either LED (light emitting diode) or MQW (multiple quantum well) laser transmitters within these intraneural links, are capable of performing satisfactorily even in the presence of such nonideal device phenomena as: 60% optical crosstalk, 50% optoelectronic device variation, or a thresholded (I(th)=0.5*I(max)) laser output characteristic. Subsequent to this, the authors then show how it is possible to use this fault tolerance to simplify the neuron architecture, to the extent that it consists only of MSM PDs a current amplifier, and an MQW laser. The overall neuron transfer function is then a first-order approximation to the original sigmoidal function.

No abstract

No abstract