Glioblastoma Classification in Hyperspectral Images by Nonlinear Unmixing

Glioblastoma is considered an aggressive tumor due to its rapid growth rate and diffuse pattern in various parts of the brain. Current in-vivo classification procedures are executed under the supervision of an expert. However, this methodology could be subjective and time-consuming. In this work, we propose a classification method for in-vivo hyperspectral brain images to identify areas affected by glioblastomas based on nonlinear spectral unmixing. This methodology follows a semi-supervised approach for the estimation of the end-members in a multilinear model. To improve the classification results, we vary the number of end-members per-class to address spectral variability of each studied type of tissue. Once the set of end-members is obtained, the classification map is generated according to the end-member with the highest abundance in each pixel, followed by morphological operations to smooth the resulting maps. The classification results demonstrate that the proposed methodology generates high performance in the regions of interest, with an accuracy above 0.75 and 0.96 in the inter and intra-patient strategies, respectively. These results indicate that the proposed methodology has the potential to be used as an assistant tool in the diagnosis of glioblastoma in hyperspectral imaging.

Automated summary

In this study, a classification method based on nonlinear spectral unmixing for identifying areas affected by glioblastomas in hyperspectral brain images is proposed. The method uses a semi-supervised approach to estimate the end-members in a multilinear model and varies the number of end-members per-class to improve classification results. The experimental results demonstrate high accuracy in the regions of interest, indicating the potential of the proposed method as an assistant tool in the diagnosis of glioblastoma in hyperspectral imaging.