Dr. Nazahah Binti Mustafa
Senior Lecturer
School of Mechatronic Engineering
Universiti Malaysia Perlis
nazahah@unimap.edu.my

NO	NAME	INSTITUTION	FACULTY/SCHOOL/ CENTRE/UNIT
1	Dr. Haniza Binti Yazid	UNIMAP	School of Mechatronic Engineering
2	Prof. Dr. Mohd Yusoff Mashor	UNIMAP	School of Mechatronic Engineering
3	Dr. Noorasmaliza Md Paiman	Kementerian Kesihatan Malaysia	Hospital Sultanah Bahiyah
4	Dr. Khairul Shakir Bin Ab Rahman	Kementerian Kesihatan Malaysia	Hospital Tuanku Fauziah

3 years (01 August 2016 – 31 July 2019)

Breast cancer is a major health concern in Malaysia. Breast cancer grading is a standard procedure used in the breast cancer diagnosis and prognosis. Manual grading of breast cancer is done using histopathological images. Manual grading is a very challenging task due to heterogeneity, low contrast, overlapping tissues and depends heavily on expertise and experience. Semi-quantitative scores obtained from manual grading are inconsistent and may lead to inter- and intra-rater variability. This study developed an automatic method to quantify breast cancer features using image processing techniques based on the Nottingham Histologic Scoring system. Features related to tubule formation, nucleus pleomorphism and mitotic activity were extracted from the breast cancer tissue images using image processing techniques such as pre-processing, segmentation, feature extraction and classification. The numerical data (i.e., the proposed quantitative measurement) obtained from the proposed image processing techniques were statistically compared with the manual scores provided by the histopathologist (semi-quantitative scores). We found that the numerical data provide a consistent way to measure the breast cancer features and could be used as a second opinion to assist histopathologists in improving breast cancer diagnosis and prognosis.

1. To fuse between segmentation and extraction techniques by modifying seed based growing algorithms to obtain breast cancer features in the region of interest.
2. To investigate structure complexity of breast cancer tissue using fractal method.
3. To validate the proposed fusion technique for measuring breast cancer features and breast cancer grading.

1. Talent:
   • 1 PHD
     1. Tan Xiao Jian (Graduated)
2. Publication:
   

   Target	Current
   6	9

   
   
   1. Article in Indexed Journals
      • Hyperchromatic Nuclei Segmentation on Breast Histopathological Images for Mitosis Detection (2018) - SCOPUS
      • Segmentation Based Classification for Mitotic Cells Detection on Breast Histopathological Images (2018) - SCOPUS
   2. Conference Proceedings
      • Segmentation Based Classification for Minimizing Number of Mitosis Candidates on Breast Histopathological Images - Perlis Research Day (2017)
      • Simple Landscapes Analysis for Relevant Regions Detection in Breast Carcinoma Histopathological Images (2018) - SCOPUS
      • Mitotic Cell Detection in Breast Histopathology Image: A Review (2019) - SCOPUS
      • Segmentation of Relevant Region in Breast Histopathology Images using FCM with Guided Initialization (2019) - SCOPUS
      • Segmentation of Irrelevant Regions Using Color Thresholding Method: Application in Breast Histopathology Images (2019) - SCOPUS
   3. Book Chapter
      • An Improved Initialization Based Histogram of K-Mean Clustering Algorithm for Hyperchromatic Nuclei Segmentation in Breast Carcinoma Histopathological Images (2018) - SCOPUS
   4. Journal Paper
      • Image Processing In Breast Carcinoma Histopathological Image: A Review (2018)

1. Human capital on medical image processing (1 Ph.D student)
2. The proposed quantitative measurement on breast cancer features could be used as a second opinion to the histopathologists.
3. The proposed quantitative measurement on breast cancer features could aid in the improvement of accuracy of the breast cancer diagnosis and prognosis and thus may lead to a reduction in the mortality rate.