Infrared Spectroscopic Imaging for Prostate Pathology

Abstract

The report summarizes progress towards using Fourier transform infrared spectroscopic imaging for prostate pathology. The aim of the work is to enable histopathologic recognition without the use of human input or stains. The major accomplishments are : 1. Optimization of experimental parameters for spectroscopic imaging experiments. The optimization provides an understanding of the classification process and allows ~40 fold decrease in data acquisition time. 2. Classification of prostate histology using Genetic Algorithms. The work explicitly identifies spectroscopic biomarkers needed to classify tissue correctly. 3. A new method is introduced to reduce noise by nearly an order of magnitude. The method is based on post-processing to reduce non-correlated signals in images using the covariance of the recorded data. The method enables a ~7 fold higher signal to noise ratio, which translates into a 50-fold faster data acquisition rate.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Mar 01, 2008
Accession Number
ADA510089

Entities

People

  • Rohit Bhargava

Organizations

  • University of Illinois Urbana–Champaign

Tags

Communities of Interest

  • Advanced Electronics

DTIC Thesaurus Topics

  • Biomedical Research
  • Brain
  • Breast Cancer
  • Carcinoma
  • Chemical Synthesis
  • Chemistry
  • Computational Science
  • Data Mining
  • Databases
  • Dimensionality Reduction
  • Factor Analysis
  • Information Processing
  • Information Science
  • Machine Learning
  • Medical Personnel
  • Supervised Machine Learning
  • Three Dimensional

Readers

  • Neural Network Machine Learning.
  • Oncology and Biomarker-Based Cancer Detection.
  • Spectroscopy.

Technology Areas

  • AI & ML
  • AI & ML - Machine Learning Algorithms
  • Biotechnology