Colorectal cancer classification based on deep ensemble model with self-adaptive training model
A Karthikeyan, S. Jothilakshmi, S. Suthir
Abstract
The four-phase colorectal cancer classification using deep ensemble model (CCCDEM) was proposed in this work. Initially, the Weiner filtering is used to preprocess the input image. The second phase produces the segmented cancer region from the preprocessed image using FCM-based segmentation. ILDTP, LGP, and MBP are only a few of the features that are extracted during the third phase. The ensemble classification model, which combines the three classifiers DBN, DMN, and CNN, is then used in the fourth phase to provide the classified output. Here, the classification outcome is determined by the optimised CNN model that trains with the scores generated from the DBN, and DMN classifiers. Since the optimal training ensures accurate categorisation, this work seeks to provide a new self-adaptive Tasmanian devil optimisation algorithm (SATDOA) for training the model by setting best weights.