Abstract/Summary

The ability to use touch in addition to vision when searching for anomalies and differences in texture is well known to be beneficial to human perception in general. The aim of this thesis is to evaluate the potential benefit of using a haptic signal in conjunction with visual images to improve detection and delineation of tumours in medical imaging data. One of the key issues with tumour delineation in the field today is the interclinician variance in delineating tumours for diagnostics and treatment, where even clinicians who have similar sensitivity and precision levels tend to delineate widely different underlying shapes. Through three experiments we investigate whether the ability to touch a medical image improves tumour delineation. In the first experiment, we show that combined visuohaptic cues significantly improves performance for signal detection of a 2D Gaussian embedded in a noisy background. In the second experiment, we found that the relative dissimilarity of different images per modality did not systematically decrease precision in a two-alternative forced choice (2AFC) slant discrimination task, in a spatially coaligned visuohaptic rig. In the third and final experiment we successfully found that observers are significantly better at delineating generated ‘tumours’ in synthetic ‘medical images’ when the haptic representation of the image is present compared to drawing on a flat surface, in a spatially coaligned visuohaptic rig.