A Resampling-Based Markovian Model for Automated Colon Cancer Diagnosis

Abstract

In recent years, there has been a great effort in the research of implementing automated diagnostic systems for tissue images. One major challenge in this implementation is to design systems that are robust to image variations. In order to meet this challenge, it is important to learn the systems on a large number of labeled images from a different range of variation. However, acquiring labeled images is quite difficult in this domain, and hence, the labeled training data are typically very limited. Although the issue of having limited labeled data is acknowledged by many researchers, it has rarely been considered in the system design. This paper successfully addresses this issue, introducing a new resampling framework to simulate variations in tissue images. This framework generates multiple sequences from an image for its representation and models them using a Markov process. Working with colon tissue images, our experiments show that this framework increases the generalization capacity of a learner by increasing the size and variation of the training data and improves the classification performance of a given image by combining the decisions obtained on its sequences.

Local Object Patterns for the Representation and Classification of Colon Tissue Images

“…Additionally, we use the resampling-based Markovian model (RMM) that we implemented in our previous work [25]. The RMM obtains multiple samples of an image, labels each sample using discrete Markov models, and votes the samples' labels to classify the image.…”

A Recent Survey on Colon Cancer Detection Techniques

“…Recently, Ozdemir et al [51] presented a resamplingbased Markovian model for classification of colon biopsy images into normal, low grade and high grade cancer. In this work, perturbed samples (images) are generated from the original image.…”

“…Table 3 demonstrates low variability in terms of accuracy. However, OO texture analysis-based techniques [44], [45], [46], [47], [48], [49], [50], [51], [52] go ahead of others. Primary reason of better accuracy is incorporation of background knowledge about tissues organization into the segmentation/classification process.…”

“…For instance, several parameter such as circle radius, window size, and merge threshold are manually adjusted in OOSEG [44]. Conversely, parameter tuning is partially automatic in others [45], [46], [47], [50], [51], [52]. In these techniques, authors have used separate data set for finding optimal values of parameters.…”

“…There are a few other techniques, called object-oriented (OO) texture analysis-based techniques, which incorporate background knowledge of normal and malignant tissue organization into the diagnostic process [44], [45], [46], [47], [48], [49], [50], [51], [52]. Focus of a few studies is to perform hyperspectral analysis on colon biopsy images, and classify samples into different classes [53], [54], [55], [56], [57], [58], [59].…”