Computer-assisted Diagnosis for Precancerous Lesions in the Esophagus

Abstract

Thus, the system could be able to support practitioners with less experience or in private practice. In combination with a connected case database it can also support case-based reasoning for the diagnostic decision process.

Predicting Pathology in Medical Decision Support Systems in Endoscopy of the Gastrointestinal Tract

“…• Number of images used While the number of patients contained in the underlying image databases is rather high in both approaches, we notice that in case of (André et al, 2009) the number of total images used is nearly four times higher compared to (Münzenmayer et al, 2009). The implies that the average number of images per patient is approximately 20 and 6 in case of (André et al, 2009) and (Münzenmayer et al, 2009), respectively.…”

“…According to our literature review 4 out of 5 methods found in total, proposed for CBIR-based prediction in case of endoscopy of the GI tract, have been published by the research group around André et al Nevertheless, we carry out a comparison of two exemplar approaches (found in André et al (2009) and Münzenmayer et al (2009)). Some details on these approaches are given in Table 2.…”

“…In contrast to (Münzenmayer et al, 2009) this is an advantage since, as already pointed out in Section 1.1, this allows in-vivo histologies. Nevertheless, André et al support their finding with a histopathologic examination.…”

“…The implies that the average number of images per patient is approximately 20 and 6 in case of (André et al, 2009) and (Münzenmayer et al, 2009), respectively.…”

“…While in (Münzenmayer et al, 2009) LOO-CV is used, André et al use LOPO-CV (the images from the patient currently under prediction are left out). Hence, compared to the experiments in (André et al, 2009), the approach validated in (Münzen-mayer et al, 2009) suffers from the fact that images from patients currently under prediction might well be contained in the training set too.…”