Spectral-Spatial Classification Using Tensor Modeling for Cancer Detection with Hyperspectral Imaging

Abstract

As an emerging technology, hyperspectral imaging (HSI) combines both the chemical specificity of spectroscopy and the spatial resolution of imaging, which may provide a non-invasive tool for cancer detection and diagnosis. Early detection of malignant lesions could improve both survival and quality of life of cancer patients. In this paper, we introduce a tensor-based computation and modeling framework for the analysis of hyperspectral images to detect head and neck cancer. The proposed classification method can distinguish between malignant tissue and healthy tissue with an average sensitivity of 96.97% and an average specificity of 91.42% in tumor-bearing mice. The hyperspectral imaging and classification technology has been demonstrated in animal models and can have many potential applications in cancer research and management.

Keywords: Dimension reduction; Feature ranking; Head and neck cancer; Hyperspectral imaging; Tensor modeling; Tucker tensor decomposition.

Figure 1

Hyperspectral imaging and illustration of pixel-based Classification

Figure 2

Spectral-spatial representation of an HSI hypercube.

Figure. 3

Flowchart of the Classification Algorithm

Figure 4

Comparison of two classification methods with different feature dimensions

Figure 5

Tumor detection results with the tensor modeling method. (a) HSI composite images of a tumor-bearing mouse. The tumor was mirrored in order to capture the whole tumor. (b) GFP composite images that serve as the gold standard for evaluation. (c) Detection results where green color represents the cancer pixels detected by the tensor modeling method.

Figure 6

Absorption spectra of cancerous and normal tissue. The left figure represents the spectra of a mouse 21 days after tumor cell injection. The right figure represents the spectra of the same mouse 35 days after injection.