Cancerous cells are found with lots of variants, some of them fall into the category of differentiated, poorly differentiated, benign and so on.
if a huge collection of pathological images are collected as data set. A data scientist may propose most accurate method to predict the cancer cell behaviors.
these images must be classified as per histologist back gourd?
is there any online system which may predict the pathological images?
If not then a collective and collaborative research may be initiated to save precious human lives and medical community may be provided additional assistance to predict the cancer cell behaviors.
Kind Regards,
JAMIL AHMED
It is an idealistic wish, that those images would be made public and data scientists could develop recognition algorithms in an open source manner. The reality is that hospitals, institutes, and sometimes patients keep the rights on this material. Imagine the monetary advantage of an institute or company holding such a technology.
But you should keep an eye on https://www.topcoder.com/?s=cancer or other platforms, they launch from time to time competitions where the goal is to develop a classification method with high accuracy. After registering, you can download image data sets.
Data-set is freely available in public domain... some of the sample sets are as follows
Ex: https://www.kaggle.com/uciml/breast-cancer-wisconsin-data
https://www.kaggle.com/gpreda/breast-cancer-prediction-from-cytopathology-data
Before proceeding with your work...you need to decide type of cancer your gonna study and based on that several public data sets can be searched in the web.
Additionally, there are already several existing methods for cancer detection through image classification technique (can be found in IEEE Explore).
Though development of such a system isn't beneficial or less likely to gain prominence among practitioner. As the clinicians often tend to not pay or opt for such tools.
I Had developed a machine learning based method in the past - "Ankush Rai, Algorithm for Early Detection of Genetic Fusion in Cancer, Tumours & Chronic Bacterial Infection, Research & Reviews: Journal of Computational Biology, Vol 3, Issue 3, 2014." Though owing to the above mentioned fact..it didn't catch the wind of commercial success.
Dear Jamil Ahmed,
Look the link, maybe useful.
Article Comprehensive analysis of lung cancer pathology images to di...
Article Machine learning applications in cancer prognosis and prediction
Regards,
Shafagat
- Badges
- Science topic