KaneAI - World’s First E2E Software Testing Agent.Book a Demo Join TestMu Conference

How to use base_image method in avocado

Best Python code snippet using avocado_python

assemble_data.py

Source:assemble_data.py

1import glob2import numpy as np3import matplotlib.pyplot as plt4from scipy import misc5import os6from tqdm import tqdm7atri_dir = 'attribute_resized/'8mask_dir = 'segmentation_resized/'9output_dir = 'semantic_map/'10file_name_arr = [] # [ISIC_00000, ISIC_000001, ISIC_000003, ...]11for file in glob.glob(atri_dir+'*.png'):12	temp = file.split('/')[-1].split('_')13	file_name = temp[0]+'_'+temp[1]14	if file_name not in file_name_arr:15	    file_name_arr.append(file_name)16if not os.path.exists(output_dir):17    os.makedirs(output_dir)18for family in tqdm(file_name_arr):19	# Create a zero filled base image20	for i, file in enumerate(glob.glob(atri_dir+family+'*.png')):21		# Read the image22		read_image = misc.imread(file, flatten=True)23		border_color = read_image[0,0]24		read_image[read_image == border_color] = 025		read_image[read_image > 0] = 25526		read_image = np.int8(read_image/255)27		if i == 0:28			mask = misc.imread(mask_dir+family+'_segmentation.png', flatten=True)29			base_image = np.ones(read_image.shape, dtype=int) # Healthy Skin is 130			border_mask_color = mask[0,0]31			base_image[mask == border_mask_color] = 032			mask[mask == border_mask_color] = 033			mask[mask > 0] = 25534			mask = np.int8(mask/255)35			base_image += mask # Common Lesion is 236		type_file = file.split('/')[-1].split('_')[3]37		if type_file == 'pigment': # 338			base_image += read_image39			if base_image[base_image > 3].any():40				base_image[base_image > 3] = 341		elif type_file == 'negative': # 442			base_image += read_image*243			if base_image[base_image > 4].any():44				base_image[base_image > 4] = 445		elif type_file.startswith('streaks'): # 546			base_image += read_image*347			if base_image[base_image > 5].any():48				base_image[base_image > 5] = 549		elif type_file == 'milia': # 650			base_image += read_image*451			if base_image[base_image > 6].any():52				base_image[base_image > 6] = 653		elif type_file.startswith('globules'): #754			base_image += read_image*555			if base_image[base_image > 7].any():56				base_image[base_image > 7] = 757		else:58			print('ERROR: Invalid File Found!!!!')...

combinators.py

Source:combinators.py

1# nuScenes dev-kit.2# Code written by Freddy Boulton, 2020.3from functools import reduce4from typing import List5import cv26import numpy as np7from nuscenes.prediction.input_representation.interface import Combinator8def add_foreground_to_image(base_image: np.ndarray,9                            foreground_image: np.ndarray) -> np.ndarray:10    """11    Overlays a foreground image on top of a base image without mixing colors. Type uint8.12    :param base_image: Image that will be the background. Type uint8.13    :param foreground_image: Image that will be the foreground.14    :return: Image Numpy array of type uint8.15    """16    if not base_image.shape == foreground_image.shape:17        raise ValueError("base_image and foreground image must have the same shape."18                         " Received {} and {}".format(base_image.shape, foreground_image.shape))19    if not (base_image.dtype == "uint8" and foreground_image.dtype == "uint8"):20        raise ValueError("base_image and foreground image must be of type 'uint8'."21                         " Received {} and {}".format(base_image.dtype, foreground_image.dtype))22    img2gray = cv2.cvtColor(foreground_image, cv2.COLOR_BGR2GRAY)23    _, mask = cv2.threshold(img2gray, 0, 255, cv2.THRESH_BINARY)24    mask_inv = cv2.bitwise_not(mask)25    img1_bg = cv2.bitwise_and(base_image, base_image, mask=mask_inv)26    img2_fg = cv2.bitwise_and(foreground_image, foreground_image, mask=mask)27    combined_image = cv2.add(img1_bg, img2_fg)28    return combined_image29class Rasterizer(Combinator):30    """31    Combines images into a three channel image.32    """33    def combine(self, data: List[np.ndarray]) -> np.ndarray:34        """35        Combine three channel images into a single image.36        :param data: List of images to combine.37        :return: Numpy array representing image (type 'uint8')38        """39        # All images in the dict are the same shape40        image_shape = data[0].shape41        base_image = np.zeros(image_shape).astype("uint8")...

augmentation.py

Source:augmentation.py

1from typing import Optional2import albumentations as A3import numpy as np4def create_green_channel(base_image: np.ndarray, noise_scale: float = 25.0, max_shift: int = 20) -> np.ndarray:5    """Data augmentation6    Args:7        base_image: (height, width)8        noise_scale (int): original image range is about 0~2559        max_shift (int):10    Returns:11        green_image: (height, width)12    """13    green_image = np.zeros_like(base_image)14    # 1. Adding gaussian noise15    base_image = base_image + np.random.randn(*base_image.shape) * noise_scale16    # base_image = np.clip(base_image, 0.0, 255.0)  # is it necessary??, maybe not.17    height, width = base_image.shape18    # 2. shift randomly, up to 2 pixel19    x_shift = np.random.randint(-max_shift, max_shift + 1)20    x_min = max(0, x_shift)21    x_max = min(width, width + x_shift)22    x_min2 = max(0, -x_shift)23    x_max2 = min(width, width - x_shift)24    y_shift = np.random.randint(-max_shift, max_shift + 1)25    y_min = max(0, y_shift)26    y_max = min(height, height + y_shift)27    y_min2 = max(0, -y_shift)28    y_max2 = min(height, height - y_shift)29    green_image[x_min:x_max, y_min:y_max] = base_image[x_min2:x_max2, y_min2:y_max2]30    return green_image31def create_green_channel_albumentations(base_image: np.ndarray, transform: Optional[A.Compose] = None) -> np.ndarray:32    """Data augmentation33    Args:34        base_image: (height, width)35        transform (None or A.Compose): albumentations transform36    Returns:37        green_image: (height, width)38    """39    if transform is None:40        # For backward compatibility41        return create_green_channel(base_image)...

Automation Testing Tutorials

Learn to execute automation testing from scratch with LambdaTest Learning Hub. Right from setting up the prerequisites to run your first automation test, to following best practices and diving deeper into advanced test scenarios. LambdaTest Learning Hubs compile a list of step-by-step guides to help you be proficient with different test automation frameworks i.e. Selenium, Cypress, TestNG etc.