Abstract:
Time-resolved photoacoustic imaging has been used to characterize breast tissue for the purpose of discriminating between normal and maligned areas of the tissue. Ultrasonic thermoelastic waves were generated in breast tissue by the absorption of nanosecond laser pulses at 193 nm produced by Q-switched Excimer laser in conjunction with a Michelson interferometer used to detect the thermoelastic waves and plot the 2-D and the 3-D image through IR- CCD. The concepts behind the use of photo-acoustic techniques for off-line detection of breast cancer tumor features were presented in earlier research papers [1][2]. This paper illustrates the application of multivariate image analysis (MIA) techniques to detect the presence of tumor features of breast cancer. MIA is used to rapidly detect the presence and quantity of common tumor features as they scanned by an RGB camera. Multiway principal component analysis is used to decompose the acquired three-channel tumor images into a two dimensional principal components (PC) space. Masking score point clusters in the score space and highlighting corresponding pixels in the image space of the two dominant PCs enables isolation of tumor defect pixels based on contrast and color information. The technique provides a qualitative result that can be used for early tumor detection. The proposed technique can potentially be used on-line to prescreen the existence of tumors through vision based systems.