Ronald Marsh

A hierarchical neural network for human face detection

This paper summarizes a research effort in human face detection. A system to locate human faces in images, especially when used as a front-end for a human face identification system, could have many applications in the law enforcement and security professions. The approach presented here is a hybrid system using an edge-enhancing preprocessor and four back-propagation neural networks arranged in a hierarchical structure. The method proposed successfully detected faces wearing glasses and all faces in images which contained multiple faces. The results obtained are reported along with a discussion for improving the system.

Fractal analysis of tumor in brain MR images

Abstract. The purpose of this study is to discuss existing fractal-based algorithms and propose novel improvements of these algorithms to identify tumors in brain magnetic-response (MR) images. Considerable research has been pursued on fractal geometry in various aspects of image analysis and pattern recognition. Magnetic-resonance images typically have a degree of noise and randomness associated with the natural random nature of structure. Thus, fractal analysis is appropriate for MR image analysis. For tumor detection, we describe existing fractal-based techniques and propose three modified algorithms using fractal analysis models. For each new method, the brain MR images are divided into a number of pieces. The first method involves thresholding the pixel intensity values; hence, we call the technique piecewise-threshold-box-counting (PTBC) method. For the subsequent methods, the intensity is treated as the third dimension. We implement the improved piecewise-modified-box-counting (PMBC) and piecewise-triangular-prism-surface-area (PTPSA) methods, respectively. With the PTBC method, we find the differences in intensity histogram and fractal dimension between normal and tumor images. Using the PMBC and PTPSA methods, we may detect and locate the tumor in the brain MR images more accurately. Thus, the novel techniques proposed herein offer satisfactory tumor identification.

Comparative analysis of de novo transcriptome assembly.

The fast development of next-generation sequencing technology presents a major computational challenge for data processing and analysis. A fast algorithm, de Bruijn graph has been successfully used for genome DNA de novo assembly; nevertheless, its performance for transcriptome assembly is unclear. In this study, we used both simulated and real RNA-Seq data, from either artificial RNA templates or human transcripts, to evaluate five de novo assemblers, ABySS, Mira, Trinity, Velvet and Oases. Of these assemblers, ABySS, Trinity, Velvet and Oases are all based on de Bruijn graph, and Mira uses an overlap graph algorithm. Various numbers of RNA short reads were selected from the External RNA Control Consortium (ERCC) data and human chromosome 22. A number of statistics were then calculated for the resulting contigs from each assembler. Each experiment was repeated multiple times to obtain the mean statistics and standard error estimate. Trinity had relative good performance for both ERCC and human data, but it may not consistently generate full length transcripts. ABySS was the fastest method but its assembly quality was low. Mira gave a good rate for mapping its contigs onto human chromosome 22, but its computational speed is not satisfactory. Our results suggest that transcript assembly remains a challenge problem for bioinformatics society. Therefore, a novel assembler is in need for assembling transcriptome data generated by next generation sequencing technique.

CONJUGATE GRADIENT AND APPROXIMATE NEWTON METHODS FOR AN OPTIMAL PROBABILISTIC NEURAL NETWORK FOR FOOD COLOR CLASSIFICATION

The probabilistic neural network (PNN) is based on the esti- mation of the probability density functions. The estimation of these den- sity functions uses smoothing parameters that represent the width of the activation functions. A two-step numerical procedure is developed for the optimization of the smoothing parameters of the PNN: a rough optimiza- tion by the conjugate gradient method and a fine optimization by the approximate Newton method. The thrust is to compare the classification performances of the improved PNN and the standard back-propagation neural network (BPNN). Comparisons are performed on a food quality problem: french fry classification into three different color classes (light, normal, and dark). The optimized PNN correctly classifies 96.19% of the test data, whereas the BPNN classifies only 93.27% of the same data. Moreover, the PNN is more stable than the BPNN with regard to the random initialization. The optimized PNN requires 1464 s for training compared to only 71 s required by the BPNN.

A fractal analysis approach to identification of tumor in brain MR images

Considerable research has been pursued on fractal geometry in various aspects of image analysis and pattern recognition. Magnetic resonance (MT) images typically have a degree of randomness associated with the natural random nature of structure. Thus fractal analysis is appropriate for MR image analysis. The purpose of this study is to apply fractal analysis to identify the presence of tumor in brain MR images. For tumor detection in MR brain images, the authors propose three different fractal-based methods. For each method, the brain MR images are divided into a number of pieces. The first method involves thresholding the pixel intensity values and hence, the authors call the technique piecewise-threshold-box-counting (PTBC) method. For the subsequent methods, the intensity is treated as the third dimension. The authors implement the improved piecewise modified box-counting (PMBC) and piecewise triangular prism surface area (PTPSA) methods respectively. With the PTBC method, they find the differences in intensity histogram and fractal dimension between normal and tumor images. Using the PMBC and PTPSA methods one can detect and locate the tumor in the brain MR images more accurately. Thus, the novel techniques proposed herein offer satisfactory tumor identification.

Assessing the Value of the OpenOrbiter Program's Research Experience for Undergraduates

This article presents an assessment of the benefits gained by undergraduate students who participated in the OpenOrbiter Small Spacecraft Development Initiative. It provides an overview of the program and its learning objectives, as they apply to undergraduate students. It compares the learning impact between students who participated and those who assumed leadership roles. Qualitative assessment with regard to benefits is also discussed. The article extrapolates from these results to identify program elements that were particularly instrumental in delivering the positive benefits discussed. Finally, future work is discussed.

Wildlife@Home: Combining Crowd Sourcing and Volunteer Computing to Analyze Avian Nesting Video

New camera technology is allowing avian ecologists to perform detailed studies of avian behavior, nesting strategies and predation in areas where it was previously impossible to gather data. Unfortunately, studies have shown mechanical triggers and a variety of sensors to be inadequate in capturing footage of small predators (e.g., snakes, rodents) or events in dense vegetation. Because of this, continuous camera recording is currently the most robust solution for avian monitoring, especially in ground nesting species. However, continuous video footage results in a data deluge, as monitoring enough nests to make biologically significant inferences results in massive amounts of data which is unclassifiable by humans alone. In the summer of 2012, Dr. Ellis-Felege gathered video footage from 63 sharp-tailed grouse (Tympanuchus phasianellus) nests, as well as preliminary interior least tern (Sternula antillarum) and piping plover (Charadrius melodus) nests, resulting in over 20,000 hours of video footage. In order to effectively analyze this video, a project combining both crowd sourcing and volunteer computing was developed, where volunteers can stream nesting video and report their observations, as well as have their computers download video for analysis by computer vision techniques. This provides a robust way to analyze the video, as user observations are validated by multiple views as well as the results of the computer vision techniques. This work provides initial results analyzing the effectiveness of the crowd sourced observations and computer vision techniques.

An overview of the OpenOrbiter autonomous operating software

The OpenOrbiter spacecraft aims to demonstrate the efficacy of the Open Prototype for Educational Nanosats (OPEN) framework. Software is an important part of this framework. This paper discusses the operating software for the spacecraft (which runs on top of the Linux operating system to command spacecraft operations). It presents an overview of this software and then pays particular attention to the aspects of software design that enable onboard autonomy. It also discusses the messaging scheme that is used onboard and the testing and validation plan. Finally, it discusses system extensibility, before concluding.

OpenOrbiter Mechanical Design: a New Approach to the Design of a 1-U CubeSat

The OpenOrbiter Small Spacecraft Development Initiative is working to create a set of designs and implementation instructions for a 1-U CubeSat, called the Open Prototype for Educational NanoSats. These designs target a total parts cost of below USD

Mining Heterogeneous ADS-B Data Sets for Probabilistic Models of Pilot Behavior

5,000. This design will be made publically available to facilitate its use by others, with or without modification. A ‘side slotted’ CubeSat design (where main circuit boards are placed in slots between the rails on the outside) has been developed for OpenOrbiter. This paper discusses the design choices that were made during the mechanical structure development of the OpenOrbiter CubeSat design, required to keep it within the mass, volume and monetary budgets. Choices like the design of the aluminum support structure, fastener mechanisms, circuit board layout and science package support structure are all discussed and their ease of construction and efficacy are considered. A discussion of ongoing work on the spacecraft’s mechanical fabrication and other subsystems is also presented. The paper also discusses how the design can, prospectively, be utilized by others and the ‘bigger picture’ benefits provided by the design approach and open hardware concept.