Robust Cell Detection of Histopathological Brain Tumor Images Using Sparse Reconstruction and Adaptive Dictionary Selection

Abstract

Successful diagnostic and prognostic stratification, treatment outcome prediction, and therapy planning depend on reproducible and accurate pathology analysis. Computer aided diagnosis (CAD) is a useful tool to help doctors make better decisions in cancer diagnosis and treatment. Accurate cell detection is often an essential prerequisite for subsequent cellular analysis. The major challenge of robust brain tumor nuclei/cell detection is to handle significant variations in cell appearance and to split touching cells. In this paper, we present an automatic cell detection framework using sparse reconstruction and adaptive dictionary learning. The main contributions of our method are: 1) A sparse reconstruction based approach to split touching cells; 2) An adaptive dictionary learning method used to handle cell appearance variations. The proposed method has been extensively tested on a data set with more than 2000 cells extracted from 32 whole slide scanned images. The automatic cell detection results are compared with the manually annotated ground truth and other state-of-the-art cell detection algorithms. The proposed method achieves the best cell detection accuracy with a F1 score = 0.96.

Performance analysis of brain tissues and tumor detection and grading system using ANFIS classifier

“…Table shows the comparisons of the proposed brain tumor segmentation methodology with conventional methodologies as Borase et al, Su et al and Evangelia et al Borase et al used K‐means classification algorithm for brain tumor segmentation which failed to detect the outlier boundary tumor pixels. The miss classification pixels in this methodology were high because of nonstability of the classes in this algorithm.…”

“…Accuracy ðAccÞ 5 ðTP1TNÞ=ðTP1FN1TN1FPÞ (9) where TP is True Positive which computes the total number of tumor pixels as tumor pixels, TN is True Negative which computes the total number of nontumor pixels as nontumor pixels, FP is False Positive which computes the total number of tumor pixels as nontumor pixels and FN is False Negative which computes the total number of nontumor pixels as tumor pixels. Table 2 shows the performance F IGUR E 8 Ground Truth images of brain tissues A, White matter; B, gray matter and C, CSF F IGUR E 9 Segmented brain tissues using proposed method A, White matter; B, gray matter and C, CSF Table 3 shows the comparisons of the proposed brain tumor segmentation methodology with conventional methodologies as Borase et al, 3 Su et al 11 and Evangelia et al 12 Borase et al 3 used K-means classification algorithm for brain tumor segmentation which failed to detect the outlier boundary tumor pixels. The miss classification pixels in this methodology were high because of nonstability of the classes in this algorithm.…”

Early detection of brain cancer using association allotment hierarchical clustering

“…To determine this, a multiple clustering approach is characterized named as hierarchical clustering which replaces single clustering algorithm to reframe different circumspection, for example, stability, lack of accuracy, and robustness. To determine the separation between the datasets and to assemble hierarchical cluster, Manhattan distance is utilized by it . Thousands of textured pictures are incorporated by an image database.…”

Detection and diagnosis of meningioma brain tumor using ANFIS classifier

“…The coefficients obtained by this transformation are trained by Support Vector Machine (SVM) classifier. Su et al 10 analyzed the cell images of various abnormal patterns in histopathological images of the brain. Alfonse et al 9 developed a brain tumor segmentation methodology using Computer Aided Detection (CAD) techniques.…”