#!/bin/python3 # ============================================================================ # Jimut Bahan Pal # May, 27, 2021 # A script to generate Mask RCNN Results # ============================================================================ import os import cv2 import json import glob import math import random import argparse import shutil import numpy as np from lxml import etree from tqdm import tqdm from PIL import Image, ImageDraw import matplotlib.pyplot as plt vibrant_colors = [[0,0,255], [0,255,0], [0,255,255], [255,0,0], [255,0,255], [255,255,0]] def parse_anno_file(cvat_xml,image_name): """ Parses annotation file and returns the details of annotation for the given image ID """ root = etree.parse(cvat_xml).getroot() # print(root) anno = [] image_name_attr = ".//image[@name='{}']".format(image_name) for image_tag in root.iterfind(image_name_attr): # print("Image tag = ",image_tag) image = {} for key, value in image_tag.items(): image[key] = value image['shapes'] = [] for poly_tag in image_tag.iter('polygon'): polygon = {'type': 'polygon'} for key, value in poly_tag.items(): polygon[key] = value image['shapes'].append(polygon) for box_tag in image_tag.iter('box'): box = {'type': 'box'} for key, value in box_tag.items(): box[key] = value box['points'] = "{0},{1};{2},{1};{2},{3};{0},{3}".format( box['xtl'], box['ytl'], box['xbr'], box['ybr']) # print("box = ",box) image['shapes'].append(box) image['shapes'].sort(key=lambda x: int(x.get('z_order', 0))) anno.append(image) return anno all_folders_root = 'Blood SmearAnalysis' all_folders = glob.glob('{}/*'.format(all_folders_root)) for folders in tqdm(all_folders): all_files_per_folder = glob.glob('{}/*'.format(folders)) all_valid_files_per_folder = [] for files in all_files_per_folder: get_extension = files.split('.')[-1] if get_extension == 'jpg' or get_extension == 'png': all_valid_files_per_folder.append(files) file_name = folders+"/annotations.xml" for valid_image_names in all_valid_files_per_folder: try: valid_image_names_ = valid_image_names.split('/')[-1] print("Image Name = ",valid_image_names_) annot = parse_anno_file(file_name,valid_image_names_) # print("valid image names_ = ",valid_image_names_) print("Annotation = ",annot) # print("--"*20) annot = annot[0] # print(json.dumps(annot, indent=4, sort_keys=True)) im_height = annot['height'] im_width = annot['width'] im_id = annot['id'] im_name = annot['name'] im_shapes = annot['shapes'] get_im_path = folders+"/"+im_name get_image = cv2.imread(get_im_path,cv2.IMREAD_COLOR) # plt.imshow(get_image[:,:,::-1]) # plt.show() # print(im_height) # print(im_width) # print(im_id) # print("Annotation name = ",im_name) name_ = im_name.split('.')[0] # read image as RGB and add alpha (transparency) im = Image.open(valid_image_names).convert("RGBA") imArray = np.asarray(im) count = 0 get_all_masks = [] get_bbox_coords = [] # x, y, w, h, label, col for shape in im_shapes: count += 1 # save_name = SAVE_FOLDER_NAME+"/"+rev_folder_map[subfolder_name]+"/"+name_+"_"+str(count)+".jpg" # print("Save Name = ",save_name) # print(shape) points = shape['points'] label = shape['label'] print(points) all_points = points.split(';') # print(all_points) x_y = [] all_x = [] all_y = [] for point_ in all_points: x = float(point_.split(',')[0]) y = float(point_.split(',')[1]) all_x.append(x) all_y.append(y) #print("X = ",x," Y = ",y) x_y.append((x,y)) # print(x_y) max_x = max(all_x) min_x = min(all_x) max_y = max(all_y) min_y = min(all_y) gap_x = max_x - min_x gap_y = max_y - min_y # print(max_x, " ", min_x, " ",max_y, " ",min_y) maskIm = Image.new('L', (imArray.shape[1], imArray.shape[0]), 0) ImageDraw.Draw(maskIm).polygon(x_y, outline=1, fill=1) # plt.imshow(maskIm) # plt.show() # ========================================= # create the mask image with certain colour maskIm = np.array(maskIm) * 255 # plt.imshow(maskIm) # plt.show() print(maskIm.max()) act_mask = np.zeros_like(get_image) act_mask[:,:,0] = maskIm act_mask[:,:,1] = maskIm act_mask[:,:,2] = maskIm print("--"*40,get_image.shape,", ",act_mask.shape) # mask_image_out = get_image.copy() green_mask = get_image.copy() col = vibrant_colors[random.randint(0,5)] green_mask[(act_mask==255).all(-1)] = col get_bbox_coords.append([min_x,min_y,gap_x,gap_y, label, col]) # plt.imshow(green_mask[:,:,::-1]) # plt.show() get_all_masks.append(green_mask) # ========================================= print("--"*40,len(get_all_masks)) final_masked_im = np.zeros_like(get_image) final_masked_im = float(1/(len(get_all_masks)+1))*get_image print("fff",final_masked_im.max()) for image in get_all_masks: print("max = ",image.max(),"min = ",image.min()) final_masked_im = final_masked_im + float(1/(len(get_all_masks)+1))*image print("Final max = ",final_masked_im.max(),"min = ",final_masked_im.min()) # plt.imshow(final_masked_im[:,:,::-1]) # plt.show() print("fin = ",final_masked_im.max()) for items in get_bbox_coords: x,y,w,h,name, col = int(items[0]), int(items[1]), int(items[2]), int(items[3]), items[4], items[5] cropped_img = np.zeros((w,h,3)) cropped_img = get_image[y:y+h,x:x+w] plt.imshow(cropped_img[:,:,::-1]) plt.show() print(x,y,w,h,name) #col = vibrant_colors[random.randint(0,5)] cv2.rectangle(final_masked_im, (x,y), (x+w,y+h), col,15) # img, text, coord, type of font, size, col, thickness cv2.putText(final_masked_im, str(name), (x, y), 0, 3, [0,0,0], 10) plt.imshow(final_masked_im[:,:,::-1]/255) plt.show() # cv2.imshow("mask", final_masked_im) # cv2.waitKey(0) except: print("PASS")