Visual explanations for polyp detection: How medical doctors assess intrinsic versus extrinsic explanations
Authors:
Steven Hicks,
Andrea Storås,
Michael Riegler,
Cise Midoglu,
Malek Hammou,
Thomas de Lange,
Sravanthi Parasa,
Pål Halvorsen,
Inga Strümke
Abstract:
Deep learning has in recent years achieved immense success in all areas of computer vision and has the potential of assisting medical doctors in analyzing visual content for disease and other abnormalities. However, the current state of deep learning is very much a black box, making medical professionals highly skeptical about integrating these methods into clinical practice. Several methods have…
▽ More
Deep learning has in recent years achieved immense success in all areas of computer vision and has the potential of assisting medical doctors in analyzing visual content for disease and other abnormalities. However, the current state of deep learning is very much a black box, making medical professionals highly skeptical about integrating these methods into clinical practice. Several methods have been proposed in order to shine some light onto these black boxes, but there is no consensus on the opinion of the medical doctors that will consume these explanations. This paper presents a study asking medical doctors about their opinion of current state-of-the-art explainable artificial intelligence methods when applied to a gastrointestinal disease detection use case. We compare two different categories of explanation methods, intrinsic and extrinsic, and gauge their opinion of the current value of these explanations. The results indicate that intrinsic explanations are preferred and that explanation.
△ Less
Submitted 23 March, 2022;
originally announced April 2022.
