Automatic deformity estimation and path tracing methodology for thoracic CT image sequences from inhale to exhale position.

Often clinical studies on image based respiratory systems either suffer road blocks or yield inconsistent results due to artifacts from a variety of subjects? erratic breathing patterns. This leads to loss of resources and time to ultimately get inconclusive and potentially wrong results. The proposed work can be seen as an automatic way of computing average thoracic deformation for a set of diverse subjects. In an image sequence, corresponding control point pairs or landmarks can be used to define the internal deformations with respect to time, point of view or modality. Defining enough number of control points in a thoracic image temporal sequence to describe the deformations happening in it is a tedious task. This inspired the use of automatic definition of control points in the proposed work. The paper proposes an automatic registration process for tracing deformity paths of the thoracic region between full inhale and exhale positions based on Hessian-matrix based feature detector and Haar wavelets based descriptor along with Optical Flow Motion (OFM) estimation technique. The proposed work presents a unique and innovative arrangement of methods to compute the average deformation of the thoracic region from all anatomical positions. The credibility and performance of the above proposed method is demonstrated by its exemplary experimental results.