This is an archive of image recognition and classification tasks from 2018
Breast cancer is a type of malignant tumor with the highest morbidity among women worldwide, which seriously endangers women's health. Besides, most breast cancers are invasive ductal carcinomas (IDC). Like a malignant tumor, breast cancers affect more than 2.5 million people worldwide. In China, 6.6% of the 2,099,000 affected individuals in 2018 died, significantly exceeding the same prevalent cancers like prostate cancer (3.8%) and colon cancer (5.8%). Because of the incidence of breast cancer among women in the eastern and central regions of China ranks first, accurately identifying and categorizing IDC and proposing treatment plans become an essential clinical task. With the development of technology, it is possible to use a computer program to determine if tumor benign or malignant with tumor section images, which can effectively decrease the probability of misdiagnosis.
In this case, it is critical that develop a program to help doctors locating the tumor and start treatment as early as possible. What our goal is developing a computer program which can identify and label breast cancer cells in the tumor section image with high accuracy.
Click me to Open/Close the directory listing
Breast cancer is a type of malignant tumor with the highest morbidity among women worldwide, which seriously endangers women's health. Besides, most breast cancers are invasive ductal carcinomas (IDC). Like a malignant tumor, breast cancers affect more than 2.5 million people worldwide. In China, 6.6% of the 2,099,000 affected individuals in 2018 died, significantly exceeding the same prevalent cancers like prostate cancer (3.8%) and colon cancer (5.8%). Because of the incidence of breast cancer among women in the eastern and central regions of China ranks first, accurately identifying and categorizing IDC and proposing treatment plans become an essential clinical task. With the development of technology, it is possible to use a computer program to determine if tumor benign or malignant with tumor section images, which can effectively decrease the probability of misdiagnosis. In this case, it is critical that develop a program to help doctors locating the tumor and start treatment as early as possible. What our goal is developing a computer program which can identify and label breast cancer cells in the tumor section image with high accuracy.
- The Breast Cancer Histopathological Annotation and Diagnosis dataset (BreCaHAD)
| Types | # of Annotations |
|---|---|
| Mitosis | 115 |
| Apoptosis | 271 |
| Tumor nuclei | 20155 |
| Non-tumor nuclei | 1905 |
| Tubule | 493 |
| Non-tubule | 610 |
| Total | 23549 |
| Tasks | Contributors |
|---|---|
| Topic Selection | Stephen Chen, Richard Liu |
| Datasets searching | Stephen Chen(Breast Cancer Locolization), Richard Liu(IDC) |
| CNN Design (Tumor Identification) | Kevin Li, Stephen Chen, Richard Liu |
| CNN Design (Tumor Locolization) | Richard Liu |
| Algorithm Optimization(Parameter) | Stephen Chen, Richard Liu, Kevin Li |
| Algorithm Optimization(Code) | Stephen Chen, Richard Liu, Kevin Li |
| Vitualization | Stephen Chen |
| Vitualization | Stephen Chen |
We would like to thank Professor Manolis Kellis for his guidance in model selection