Data Analysis of X-Ray Image System (DAXIS), Phase 1.
Abstract
The AnSim project DAXIS - Data Analysis of X-Ray Image System, developed a prototype digital imaging processing sequence for object identification and classification using the filtering/predictive capabilities of wavelets, fractals and artificial neural networks (ANN). Each of these analytical techniques has its own characteristics which, when used in combination, were anticipated to synergistically assist in meeting the overall objective. A number of wavelet families were investigated as image filters for object location and identification. When combined with conventional imaging processing techniques and a fractal dimension calculation, a number of quantitative object characteristics were determined. These characteristics were placed in a feature vector and submitted to neural network analysis for purposes of object classification. The prototype was developed and a set of x-rays were used to train a neural network which was then tested in classifying anomalies found in another government-furnished set of x-rays.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 15, 1997
- Accession Number
- ADA327541
Entities
People
- James A. Leone
- Jingquan Li
- Paul Patti
- Piara Arzumand
- Xinran Wang