Localization of the vertebrae is importance in many medical applications. For example, can serve as landmarks image registration. They also provide a reference coordinate system to facilitate localization other organs chest. In this paper, we propose new method using convolutional neural networks (CNN). The main advantage proposed removal hand-crafted features. We construct two training sets train CNNs that share same architecture. One used distinguish from tissues chest, and aimed at...

In order to extract the low contrast vessels in retinal images, we propose a novel enhancement algorithm based on directional field and develop robust vessel extraction framework. The proposed can significantly improve of vasculature especially slim vessels. our framework, first enhance by algorithm, from enhanced image local method. Secondly, apply series modified top-hat transform original image, use binary morphological reconstruction large size Finally, whole is provided an organic...

Moldy cores in apples are not initially obvious from the outside of fruit, so developing methods to detect moldy is an important area research apple industry. The objective this study was improve ability near-infrared spectrometry apples. Transmission spectra were recorded for 200 samples range 200–1100[Formula: see text]nm, and 140 60 randomly selected as training test sets, respectively. Signal de-noising performed by wavelet thresholding based on results orthogonal experiments. best...

Current certificate-based information security technologies are facing a challenge of lacking the exact connection between cryptographic key and legitimate users problem comprehensive management certificates. In this paper we propose novel networking identity authentication scheme based on fingerprint coding method Identity-based encryption (IBE). Fingerprint code is dual-factor authenticator which combines token, e.g. USB key, with user's feature by direct mixing pseudo-random number...

It is difficult to determine which apples have moldy cores just by looking at the outside of apple. In present study, we investigated identifying using near-infrared transmittance spectra. First, input spectral features selected noise adjusted principal component analysis (NAPCA) for back propagation artificial neural network (BP ANN) was used reduce dimensions original data. Then, four factors and five levels uniform design nodes, training functions, transfer layer functions output NAPCA-BP...

In order to improve the segmentation accuracy of plant lesion images, multi-channels algorithm disease image was proposed based on linear discriminant analysis (LDA) method’s mapping and K-means’ clustering. Firstly, six color channels from RGB model HSV were obtained, all pixels laid out columns. Then one these regarded as label others sample features. These data grouped for discrimination analysis, values other five applied eigen vector space according first three big values. Secondly,...

Various types of advanced imaging technologies have significantly improved the quality medical care available to patients. Corresponding image reconstruction algorithms, especially 3D reconstruction, play an important role in disease diagnosis and treatment assessment. However, these increasing methods are not implemented a unified software framework, which brings along lots disadvantages such as breaking connection different modalities, lack module reuse inconvenience method comparison....

Fingerprint recognition is implemented on Ti DSP chip 5402. The algorithm prompt and practicable. Enhance part adopts a fast orientation filter to improve inputting image quality combining interpolation method. On the other hand, field used in fingerprint segment. Matching uses point match method combines some post-processing order get good result. One fusion way introduced enrollment. In implement phase, system frame one kind of dispatched structure aiming at 5402 features. Otherwise...

A fast surface reconstruction algorithm is proposed in this paper for processing large scale industrial CT images with high resolution. Through the following main steps: tracking single layer, a data caching mechanism and triangle strip generation, can extract represent surfaces efficiently on current mainstream PCs. The experimental results tested real datasets are also reported.