This repository is for the image processing portion of our early detection of breast cancer using machine learning
Command: python image_preproc.py 'results.csv'
Description: Uses the txt files from './patientsAssignments/' folder to discover our cancerous vs non-cancerous
sets. After splitting into the sets, it reads the processPatient function data into result arrays.
(If desired) You can enter a filename like 'results.csv' in the command to save the output to a csv file.
Description: Standalone function built to return a vector of image metadata for our ensemble classifier. The input parameter should be an image file, preferably read in by OpenCV's cv2.imread function. The function computes a series of statistics documented below.
Parameter: im - image file read in by cv2.imread(img,1)
Return: List < Integer > - Data from the patient
List[0 - 2]: skew of r,g,b
List[3-5]: kurtosis of r,g,b
List[6-8]: total image mean pixel r,g,b values
List[9-11]: left half of image mean pixel r,g,b
List[12-14]: right half of image mean pixel r,g,b
Parameter: f_im { cv2Image } - f version of the front facing patient image
Parameter: fc_im { cv2 Image } - fc version of the front facing patient image
Return: { List } - concatenation of the processing of f and fc image lists