Gaze2segment: a pilot study for integrating eye-tracking technology into medical image segmentation
Medical Computer Vision and Bayesian and Graphical Models for Biomedical Imaging: MICCAI 2016 International Workshops, MCV and BAMBI, Athens, Greece, October 21, 2016, Revised Selected Papers, pp. 94–104
Abstract
This study introduced a novel system, called Gaze2Segment, integrating biological and computer vision techniques to support radiologists' reading experience with an automatic image segmentation task. During diagnostic assessment of lung CT scans, the radiologists' gaze information were used to create a visual attention map. This map was then combined with a computer-derived saliency map, extracted from the gray-scale CT images. The visual attention map was used as an input for indicating roughly the location of a object of interest. With computer-derived saliency information, on the other hand, we aimed at finding foreground and background cues for the object of interest. At the final step, these cues were used to initiate a seed-based delineation process. Segmentation accuracy of the proposed Gaze2Segment was found to be 86% with dice similarity coefficient and 1.45 mm with Hausdorff distance. To the best of our knowledge, Gaze2Segment is the first true integration of eye-tracking technology into a medical image segmentation task without the need for any further user-interaction.
Abstract sourceThis 2017 workshop-proceedings paper presents a pilot system that combines radiologists’ eye gaze with image-derived cues to guide segmentation of lung lesions in chest CT scans. The evaluation used four CT volumes; the authors identify the study’s limited data and other constraints.
Summary
Gaze2Segment combines radiologists’ eye-tracking data with image-derived saliency and gradient cues to identify regions of interest and initialize seed-based segmentation. The authors evaluated the system on four chest CT volumes, read by three radiologists. They report an average Dice similarity coefficient of 86% and an average Hausdorff distance of 1.45 mm. These are pilot-study results, not evidence of clinical readiness.
The authors note that the evaluation used a limited number of scans, that some non-lesion regions were also identified, and that segmentation was performed offline. They describe online processing and evaluation on additional data as future work.
- Naji Khosravan, Haydar Celik, Baris Turkbey, Ruida Cheng, Evan McCreedy, Matthew McAuliffe, Sandra Bednarova, Elizabeth Jones, Xinjian Chen, Peter L. Choyke, Bradford J. Wood, Ulas Bagci.
Cite this work
Naji Khosravan, Haydar Celik, Baris Turkbey, Ruida Cheng, Evan McCreedy, Matthew McAuliffe, Sandra Bednarova, Elizabeth Jones, Xinjian Chen, Peter L. Choyke, Bradford J. Wood, Ulas Bagci (2017). Gaze2segment: a pilot study for integrating eye-tracking technology into medical image segmentation. Medical Computer Vision and Bayesian and Graphical Models for Biomedical Imaging: MICCAI 2016 International Workshops, MCV and BAMBI, Athens, Greece, October 21, 2016, Revised Selected Papers, pp. 94–104 https://doi.org/10.1007/978-3-319-61188-4_9
BibTeX