Extracting features from tactile maps

Abstract

Tactile imaging is a newly developed mechanical sensing technology for documenting the properties of hard lumps contained in soft tissue. An examiner strokes a scan head across tissue that contains a mass and images of the distributed contact pressure between the head and the tissue are recorded. We have developed models that predict these pressure distributions from geometric and material properties. We then use inversion algorithms developed from these models to extract lump size and shape. In a limited clinical trial on 24 surgical patients, lump size was estimated with less than 17% mean absolute error when compared with ex-vivo size measurements. This is more than twice as accurate as either clinical breast examination or ultrasound examination of the same lumps. This result demonstrates that tactile imaging has the potential to improve the accuracy of clinical breast examination.