Colonoscopy is the gold standard for colon cancer screening though some polyps are still missed, thus preventing early disease detection and treatment. Several computational systems have been proposed to assist polyp detection during colonoscopy but so far without consistent evaluation. The lack of publicly available annotated databases has made it difficult to compare methods and to assess if they achieve performance levels acceptable for clinical use. The Automatic Polyp Detection sub-challenge, conducted as part of the Endoscopic Vision Challenge (http://endovis.grand-challenge.org) at the international conference on Medical Image Computing and Computer Assisted Intervention (MICCAI) in 2015, was an effort to address this need. In this paper, we report the results of this comparative evaluation of polyp detection methods, as well as describe additional experiments to further explore differences between methods. We define performance metrics and provide evaluation databases that allow comparison of multiple methodologies. Results show that convolutional neural networks are the state of the art. Nevertheless, it is also demonstrated that combining different methodologies can lead to an improved overall performance.

Background and aims: Polyp miss-rate is a drawback of colonoscopy that increases significantly for small polyps. We explored the efficacy of an automatic computer-vision method for polyp detection. Methods: Our method relies on a model that defines polyp boundaries as valleys of image intensity. Valley information is integrated into energy maps that represent the likelihood of the presence of a polyp. Results: In 24 videos containing polyps from routine colonoscopies, all polyps were detected in at least one frame. The mean of the maximum values on the energy map was higher for frames with polyps than without (P < 0.001). Performance improved in high quality frames (AUC = 0.79 [95 %CI 0.70 – 0.87] vs. 0.75 [95 %CI 0.66 – 0.83]). With 3.75 set as the maximum threshold value, sensitivity and specificity for the detection of polyps were 70.4 % (95 %CI 60.3 % – 80.8 %) and 72.4 % (95 %CI 61.6 % – 84.6 %), respectively. Conclusion: Energy maps performed well for colonic polyp detection, indicating their potential applicability in clinical practice.