import sys import os import pandas as pd import numpy as np from PIL import Image, ImageDraw, ImageFilter import glob import cv2 SOURCEDIR = os.path.join('..', 'input') # ================ Meta ==================== __description__ = 'Sea Lion Prognostication Engine' __version__ = '0.1.0' __license__ = 'MIT' __author__ = 'Gavin Crooks (@threeplusone)' __status__ = "Prototype" __copyright__ = "Copyright 2017" class SeaLionData(object): def __init__(self, sourcedir=SOURCEDIR): self.class_names = ( 'adult_males', 'subadult_males', 'adult_females', 'juveniles', 'pups') self.class_colors = ( (255,0,0), # red (250,10,250), # magenta (84,42,0), # brown (30,60,180), # blue (35,180,20), # green ) self._trainshort_nb = 11 self.train_nb = 947 self.test_nb = 18636 self.source_paths = { 'sample' : os.path.join(sourcedir, 'sample_submission.csv'), 'counts' : os.path.join(sourcedir, 'Train', 'train.csv'), 'train' : os.path.join(sourcedir, 'Train', '{iid}.jpg'), 'dotted' : os.path.join(sourcedir, 'TrainDotted', '{iid}.jpg'), 'test' : os.path.join(sourcedir, 'Test', '{iid}.jpg'), } # From MismatchedTrainImages.txt self.bad_train_ids = ( 3, 7, 9, 21, 30, 34, 71, 81, 89, 97, 151, 184, 215, 234, 242, 268, 290, 311, 331, 344, 380, 384, 406, 421, 469, 475, 490, 499, 507, 530, 531, 605, 607, 614, 621, 638, 644, 687, 712, 721, 767, 779, 781, 794, 800, 811, 839, 840, 869, 882, 901, 903, 905, 909, 913, 927, 946) def class_from_color(self, rgb) : MAX_DIFF = 80 colors = np.array(self.class_colors) diff = (abs(colors - rgb)).sum(axis=-1) cls = np.argmin(diff) if diff[cls]>MAX_DIFF: #print(cls, rgb, self.class_colors[cls], diff[cls]) return None return cls @property def trainshort_ids(self): return (0,1,2,4,5,6,8,10) #range(41,51) @property def train_ids(self): tids = range(0, self.train_nb) tids = list(set( tids) - set(self.bad_train_ids) ) # Remove bad ids tids.sort() return tids @property def test_ids(self): return range(0, self.test_nb) def path(self, name, **kwargs): """Return path to various source files""" path = self.source_paths[name].format(**kwargs) return path def counts(self) : counts = {} fn = self.path('counts') with open(fn) as f: f.readline() for line in f: tid_counts = list(map(int, line.split(','))) counts[tid_counts[0]] = tid_counts[1:] return counts def load_train_image(self, train_id): fn = self.path('train', iid=train_id) img = Image.open(fn) return img def load_dotted_image(self, train_id): fn = self.path('dotted', iid=train_id) img = Image.open(fn) return img def coords(self, train_id): # Empirical constants MIN_DIFFERENCE = 30 MIN_AREA = 15 MAX_AREA = 100 MAX_AVG_DIFF = 50 ''' src_fn = self.path('train', iid=train_id) src_img = np.asarray(Image.open(src_fn), dtype = np.float) print(cv2.mean(src_img)) print(src_img.shape) dot_fn = self.path('dotted', iid=train_id) dot_img = np.asarray(Image.open(dot_fn), dtype = np.float) print(cv2.mean(dot_img)) img_diff = np.abs(src_img-dot_img) print(cv2.mean(img_diff)) print(img_diff.shape) # Detect bad data. If train and dotted images very different somethings wrong. img_diff[dot_img==0] = 0 # Mask out black masks in dotted images avg_diff = img_diff.sum() / (img_diff.shape[0] * img_diff.shape[1]) print(avg_diff) ''' src_fn = self.path('train', iid=train_id) src_img = cv2.cvtColor(cv2.imread(src_fn),cv2.COLOR_BGR2RGB) dot_fn = self.path('dotted', iid=train_id) dot_img = cv2.cvtColor(cv2.imread(dot_fn),cv2.COLOR_BGR2RGB) img_diff = cv2.absdiff(dot_img,src_img) img_diff[dot_img==0] = 0 # Mask out black masks in dotted images avg_diff = img_diff.sum() / (img_diff.shape[0] * img_diff.shape[1]) if avg_diff > MAX_AVG_DIFF: return None img_diff[dot_img==0] = 0 # Mask out black masks in dotted images img_diff = np.max(img_diff, axis=-1) img_diff[img_diff=MIN_DIFFERENCE] = 255 # debug #img = Image.fromarray(img_diff.astype(np.uint8) ) #img.save('img_diff.png') sealions = [] contours = cv2.findContours(img_diff.astype(np.uint8), cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE)[-2] for cnt in contours : area = cv2.contourArea(cnt) if(area>MIN_AREA and area