Classifying breast cancer regions in microscopic image using texture analysis and neural network

Cell type classifiers for breast cancer microscopic images based on fractal dimension texture analysis of image color layers

“…This is caused by the unappropriated selection of color channel. Moreover, while four combination features and a neural network classifier showed 94.23% accuracy in the previous results (Jitaree et al ., ), the single feature from the FD value of Cr color channel can provide comparable accuracy at 93.87% evaluated from 800 images with the same window size.…”

“…These results show that the appropriate selection of color channel is very important and has a significant effect on the accuracy. Compared with the previous results, performance evaluation from 104 images with window size 256 gave an accuracy of 68.27% only (Jitaree et al ., ). This is caused by the unappropriated selection of color channel.…”

“…The FD value determined from binary box‐counting method together with other three features consisting of (1) the correlation; (2) auto‐correlation; and (3) information measure of correlation determined from the gray level co‐occurrence matrices were used to classify BCMI into three categories, i.e SC, LC, and cancer cells (Jitaree et al ., ). Performance evaluated from 104 images of four combination features and a neural network classifier showed 94.23% accuracy.…”

Mining textural knowledge in biological images: Applications, methods and trends

“…Support Vector Machines [28], [29], boosting algorithms [30], [31], neural networks [32], [33], and random forest techniques [34], [35]), most of the efforts of the research community are directed towards designing suitable texture descriptors for specific biological applications. Indeed, literature suggests that a smart choice and encoding of the features is by far the most important aspect in obtaining a accurate texture discrimination [36].…”

Automatic classification of histological images and histological knowledge modelling of the human cardivascular system = Clasificación automática de imágenes histológicas y modelado del conocimiento histológico del sistema cardiovascular humano

“…However, this technique encounters several problems due to the low definition in some edges of the tissues and organs (Alén et al, 2006;Fujii et al, 2013;Fakhrzadeh et al, 2013). On the other hand, classification and recognition process are based on: (i) clustering methods such as K-means, mean shift, K-nn, among others are employed for histological image segmentation (Wu et al, 2015;He et al, 2011); (ii) SVM, Bayesian networks and another machine learning algorithms are used to characterise and classify healthy and pathology cells, tissues and organs (d. A. Zampirolli et al, 2010;Krishnan et al, 2010;Veillard et al, 2012); (iii) neural networks are used to the classification task producing good results on automatically learned features (Bevilacqua et al, 2015;Jitaree et al, 2013;Kashif et al, 2016); (iv) learning based approach using bag-of-words model, this method is attractive as it offers good classification accuracy at low computation cost than the texture-feature-based methods (Galaro et al, 2011;Nguyen et al, 2015;Cheng et al, 2012). However, some of these works are focus in histopathologic images which have different characteristics to healthy tissues.…”