#Full Research is here :
https://ieeexplore.ieee.org/document/8621307
Breast Cancer Histopathological Image Classification: A Deep Learning Approach
Breast cancer remains the most common type of cancer and the leading cause of cancer-induced mortality among women with 2.4 million new cases diagnosed and 523, 000 deaths per year. Historically, a diagnosis has been initially performed using clinical screening followed by histopathological analysis. Automated classification of cancers using histopathological images is a challenging task of accurate detection of tumor sub-types. This process could be facilitated by machine learning approaches, which may be more reliable and economical compared to conventional method.
To prove this principle, we applied fine-tuned pre-trained deep neural networks and first attempted to discriminate be- tween different cancer types. Using 6, 402 tissue microarrays (TMAs) samples, models including the ResNet V1 50 pre- trained model correctly predicted 99.8% of the four cancer types including breast, bladder, lung, and lymphoma. Then, for classification of breast cancer sub-types, this approach was applied to 7, 909 images of 82 patients from the BreakHis database. ResNet V1 152 classified benign and malignant breast cancers with an accuracy of 98.7%. In addition, ResNet V1 50 and ResNet V1 152 categorized either benign- (adeno-sis, fibroadenoma, phyllodes tumor, and tubular adenoma) or malignant- (ductal carcinoma, lobular carcinoma, mucinous carcinoma, and papillary carcinoma) sub-types with 94.8% and 96.4% accuracy, respectively. The confusion matrices revealed high sensitivity values of 1, 0.995 and 0.993 for cancer types, as well as malignant- and benign sub-types respectively. The areas under the curve (AUC) scores were 0.996, 0.973 and 0.996 for cancer types, malignant and benign sub-types, respectively. One of the most significant and striking result to emerge from the data analysis is negligible false positive (FP) and false negative (FN). The optimum results, as shown in Tables, indicate that FP is between 0 and 4 while FN is between 0 and 8 in which test data including 800, 900, 809, 1000 for given four classes
Python 3+
Platform Linux Ubuntu 16.04
NVIDIA GPU
Tensorflow-gpu 1.7+ and its dependencies ( https://www.tensorflow.org/install/gpu)
Extract all files and keep that folders name as it is.
Gentle reminder: 4.TFslim_fine_tune.zip should be Extracted and kept as a folder with its own contents
guidline.pdf is a very early basic guide (Version 1 is incomplete, But ship it anyway)
This work has been supported in part by a start-up fund from Weill Cornell Medicine and the Iranian National Elite Foundation and grants provided by Royan Institute.
If you find this code/ research useful in your research, please consider citing:
Plain Text : M. Jannesari et al., "Breast Cancer Histopathological Image Classification: A Deep Learning Approach," 2018 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), Madrid, Spain, 2018, pp. 2405-2412.
BibTex:
@INPROCEEDINGS{BreastCancerBIBM2018, author={M. Jannesari and M. Habibzadeh and H. Aboulkheyr and P. Khosravi and O. Elemento and M. Totonchi and I. Hajirasouliha}, booktitle={2018 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)}, title={Breast Cancer Histopathological Image Classification: A Deep Learning Approach}, year={2018}, volume={}, number={}, pages={2405-2412}, keywords={Breast cancer;Deep learning;Tumors;Pathology;Training;Convolutional Neural Network;Deep learning;Digital Pathology Imaging;Breast cancer}, doi={10.1109/BIBM.2018.8621307}, ISSN={}, month={Dec},}
Mahboobeh Jannesari ([email protected])
Mehdi Habibzadeh ([email protected])
Hamidreza AB ES ([email protected])
Mehdi Totonchi ([email protected])
Iman Hajirasouliha ([email protected])
Olivier Elemento ([email protected])