Infrared hyperspectral imaging for enhanced tissue visualization and discrimination during surgical operation

A new hyperspectral imaging system has been designed in infrared 950 nm-1700 nm spectral range to be integrated in operating rooms in order to detect some vital and anatomical tissues, hardly noticeable by the surgeon´s naked eye during surgical interventions. The main objective of this exploratory study is to find the best three relevant bands to create a RGB image that maximizes the contrast between the target tissue and its surroundings. One of the main challenges in HSI processing is the dimensionality reduction step which was carried out using band transformation and band selection methods in order to remove redundancy in the collected data without losing valuable information. Preclinical experiments were performed in vivo in a porcine ureter. Objective and comparative evaluations, based on the simulated perceived image, were performed between five methods. Experimental results show that the proposed method that combines two band selection methods has the best performance to enhance surgeon´s visual skills.