# -*- coding: utf-8 -*- """DIS.ipynb Automatically generated by Colaboratory. Original file is located at https://colab.research.google.com/drive/1MI9utM7GJbz0w_zw1GJNU-ay15SzZHIN # Clone official repo """ # Commented out IPython magic to ensure Python compatibility. ! git clone https://github.com/xuebinqin/DIS # %cd ./DIS/IS-Net !pip install gdown !mkdir ./saved_models """# Imports""" import numpy as np from PIL import Image import torch from torch.autograd import Variable from torchvision import transforms import torch.nn.functional as F import gdown import os import requests import matplotlib.pyplot as plt from io import BytesIO # project imports from data_loader_cache import normalize, im_reader, im_preprocess from models import * """# Helpers""" drive_link = "https://drive.google.com/uc?id=1XHIzgTzY5BQHw140EDIgwIb53K659ENH" # Specify the local path and filename local_path = "/content/DIS/IS-Net/saved_models/isnet.pth" # Download the file gdown.download(drive_link, local_path, quiet=False) device = 'cuda' if torch.cuda.is_available() else 'cpu' # Download official weights class GOSNormalize(object): ''' Normalize the Image using torch.transforms ''' def __init__(self, mean=[0.485,0.456,0.406], std=[0.229,0.224,0.225]): self.mean = mean self.std = std def __call__(self,image): image = normalize(image,self.mean,self.std) return image transform = transforms.Compose([GOSNormalize([0.5,0.5,0.5],[1.0,1.0,1.0])]) def load_image(im_path, hypar): if im_path.startswith("http"): im_path = BytesIO(requests.get(im_path).content) im = im_reader(im_path) im, im_shp = im_preprocess(im, hypar["cache_size"]) im = torch.divide(im,255.0) shape = torch.from_numpy(np.array(im_shp)) return transform(im).unsqueeze(0), shape.unsqueeze(0) # make a batch of image, shape def build_model(hypar,device): net = hypar["model"]#GOSNETINC(3,1) # convert to half precision if(hypar["model_digit"]=="half"): net.half() for layer in net.modules(): if isinstance(layer, nn.BatchNorm2d): layer.float() net.to(device) if(hypar["restore_model"]!=""): net.load_state_dict(torch.load(hypar["model_path"]+"/"+hypar["restore_model"],map_location=device)) net.to(device) net.eval() return net def predict(net, inputs_val, shapes_val, hypar, device): ''' Given an Image, predict the mask ''' net.eval() if(hypar["model_digit"]=="full"): inputs_val = inputs_val.type(torch.FloatTensor) else: inputs_val = inputs_val.type(torch.HalfTensor) inputs_val_v = Variable(inputs_val, requires_grad=False).to(device) # wrap inputs in Variable ds_val = net(inputs_val_v)[0] # list of 6 results pred_val = ds_val[0][0,:,:,:] # B x 1 x H x W # we want the first one which is the most accurate prediction ## recover the prediction spatial size to the orignal image size pred_val = torch.squeeze(F.upsample(torch.unsqueeze(pred_val,0),(shapes_val[0][0],shapes_val[0][1]),mode='bilinear')) ma = torch.max(pred_val) mi = torch.min(pred_val) pred_val = (pred_val-mi)/(ma-mi) # max = 1 if device == 'cuda': torch.cuda.empty_cache() return (pred_val.detach().cpu().numpy()*255).astype(np.uint8) # it is the mask we need """# Set Parameters""" hypar = {} # paramters for inferencing hypar["model_path"] ="./saved_models" ## load trained weights from this path hypar["restore_model"] = "isnet.pth" ## name of the to-be-loaded weights hypar["interm_sup"] = False ## indicate if activate intermediate feature supervision ## choose floating point accuracy -- hypar["model_digit"] = "full" ## indicates "half" or "full" accuracy of float number hypar["seed"] = 0 hypar["cache_size"] = [1024, 1024] ## cached input spatial resolution, can be configured into different size ## data augmentation parameters --- hypar["input_size"] = [1024, 1024] ## mdoel input spatial size, usually use the same value hypar["cache_size"], which means we don't further resize the images hypar["crop_size"] = [1024, 1024] ## random crop size from the input, it is usually set as smaller than hypar["cache_size"], e.g., [920,920] for data augmentation hypar["model"] = ISNetDIS() """# Build Model""" net = build_model(hypar, device) """# Predict Mask""" gsheetid = "1n9kk7IHyBzkw5e08wpjjt-Ry5aE_thqGrJ97rMeN-K4" sheet_name = "sarvm" gsheet_url = "https://docs.google.com/spreadsheets/d/{}/gviz/tq?tqx=out:csv&sheet={}".format(gsheetid, sheet_name) gsheet_url import pandas as pd df = pd.read_csv(gsheet_url) image_path = df.iloc[-1]['Image'] drive_link = image_path # Specify the local path and filename local_path = "/content/DIS/IS-Net/saved_models/input2.jpg" # Download the file gdown.download(drive_link, local_path, quiet=False) from google.colab.patches import cv2_imshow from PIL import Image image_path = "/content/DIS/IS-Net/saved_models/input2.jpg" # image_bytes = BytesIO(requests.get(image_path).content) # print(image_bytes) image_tensor, orig_size = load_image(image_path, hypar) mask = predict(net,image_tensor,orig_size, hypar, device) image = Image.open(image_path) f, ax = plt.subplots(1,2, figsize = (35,20)) # ax[0].imshow(np.array(Image.open(image_bytes))) # Original image # cv2_imshow(image_path) ax[0].imshow(mask, cmap = 'gray') # retouched image # ax[0].set_title("Original Image") ax[0].set_title("Mask") plt.show() import cv2 image = cv2.imread(image_path) h, w , _ = image.shape # print(h) # print(w) # print(_) # print(image) h, w , _ = image.shape # print(h) # print(w) # print(_) # new_image = np.zeros_like(image) # new_image[mask] = image[mask] new_image = cv2.bitwise_and(image, image, mask=mask) transparent_bg = np.zeros((new_image.shape[0],new_image.shape[1], new_image.shape[2]+1) , dtype=np.uint8) # Apply the mask to the transparent background transparent_bg[:, :, :3] = new_image # Set the alpha channel using the mask transparent_bg[:, :, 3] = mask # Save the new image with a transparent background output_path = "/content/output.png" cv2.imwrite(output_path, transparent_bg) # Save the new image # output_path = "/content/output.jpg" # cv2.imwrite(output_path, new_image)