Feature/white background

app/Colab.ipynb

@@ -0,0 +1 @@
+{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"provenance":[],"authorship_tag":"ABX9TyNzifwmIaQmmM75iyCsAofz"},"kernelspec":{"name":"python3","display_name":"Python 3"},"language_info":{"name":"python"},"gpuClass":"standard"},"cells":[{"cell_type":"code","source":["from google.colab import drive\n","drive.mount('/content/drive')\n","\n","!ls /content/drive/"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"yur7vTYv3wxx","executionInfo":{"status":"ok","timestamp":1697182836766,"user_tz":-540,"elapsed":3062,"user":{"displayName":"威廷旼","userId":"02776276979241908427"}},"outputId":"dd210930-6f4e-45d2-abea-5c6670236ac0"},"execution_count":11,"outputs":[{"output_type":"stream","name":"stdout","text":["Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n","MyDrive\n"]}]},{"cell_type":"code","source":["!ls -l /content/drive/MyDrive/"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"48ZCoYhL4BwC","executionInfo":{"status":"ok","timestamp":1697182839001,"user_tz":-540,"elapsed":366,"user":{"displayName":"威廷旼","userId":"02776276979241908427"}},"outputId":"77152fc3-1eca-42a7-83b2-fc4ff6bb681d"},"execution_count":12,"outputs":[{"output_type":"stream","name":"stdout","text":["total 1377\n","-rw------- 1 root root 1392695 Aug 18 05:38 a.xlsx\n","drwx------ 2 root root    4096 Feb 20  2022 Classroom\n","drwx------ 4 root root    4096 Oct 13 06:25 DeepL\n","drwx------ 2 root root    4096 Nov 11  2022 RCSC\n","drwx------ 2 root root    4096 Oct 25  2022 SEAGULL\n","-rw------- 1 root root     171 Jul 31 05:35 無題のドキュメント.gdoc\n"]}]},{"cell_type":"code","source":["!ls -l /content/drive/MyDrive/DeepL/Img-Bg-White"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"XOaTC_k34Dzj","executionInfo":{"status":"ok","timestamp":1697182842377,"user_tz":-540,"elapsed":358,"user":{"displayName":"威廷旼","userId":"02776276979241908427"}},"outputId":"5f5c3f74-6cc5-4929-afdc-ccae07f23897"},"execution_count":13,"outputs":[{"output_type":"stream","name":"stdout","text":["total 93\n","-rw------- 1 root root 57878 Oct 13 07:35 1.jpg\n","-rw------- 1 root root     0 May 16 04:33 app.py\n","-rw------- 1 root root   147 May 16 04:33 config.py\n","-rw------- 1 root root  8772 May 16 04:33 data_loader.py\n","-rw------- 1 root root  3502 Oct 13 07:40 __init__.py\n","drwx------ 3 root root  4096 Oct 13 06:30 model\n","-rw------- 1 root root  1035 May 16 04:33 README.md\n","-rw------- 1 root root   480 May 16 04:33 requirements.txt\n","drwx------ 3 root root  4096 Oct 13 06:30 saved_models\n","drwx------ 5 root root  4096 Oct 13 06:30 static\n","-rw------- 1 root root  8447 Oct 13 07:36 Untitleaaa.ipynb\n"]}]},{"cell_type":"code","source":["%cd /content/drive/MyDrive/DeepL/Img-Bg-White"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"IwJvGc6y69X1","executionInfo":{"status":"ok","timestamp":1697182845101,"user_tz":-540,"elapsed":2,"user":{"displayName":"威廷旼","userId":"02776276979241908427"}},"outputId":"99819aff-14a2-408f-b2fb-ff70d778ee93"},"execution_count":14,"outputs":[{"output_type":"stream","name":"stdout","text":["/content/drive/MyDrive/DeepL/Img-Bg-White\n"]}]},{"cell_type":"code","source":["%run __init__.py"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"SqCHlLVf4Rqp","executionInfo":{"status":"ok","timestamp":1697182856287,"user_tz":-540,"elapsed":7954,"user":{"displayName":"威廷旼","userId":"02776276979241908427"}},"outputId":"ca0c975d-1211-489d-9976-43001b43d4a4"},"execution_count":15,"outputs":[{"output_type":"stream","name":"stdout","text":["---Loading Model---\n","---Removing Background...\n","---Success---\n"]}]},{"cell_type":"markdown","source":["THIS ONE WILL NOT WORK"],"metadata":{"id":"iD49WxAW9HDl"}},{"cell_type":"code","source":["%mkdir saved_models\n","%mkdir saved_models/u2net"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"cUpYjA5Q_F3d","executionInfo":{"status":"ok","timestamp":1684214153158,"user_tz":-540,"elapsed":1142,"user":{"displayName":"威廷旼","userId":"02776276979241908427"}},"outputId":"ae8c0e6b-64e4-4151-fbc6-eaf3a4485877"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["mkdir: cannot create directory ‘saved_models’: File exists\n","mkdir: cannot create directory ‘saved_models/u2net’: File exists\n"]}]},{"cell_type":"code","source":["path_proj = '/content/drive/MyDrive/DeepDark/Img-Bg-Remover-Wk'\n","\n","with open(path_proj) as f:\n"],"metadata":{"id":"yUXO0cOX5Jlq"},"execution_count":null,"outputs":[]},{"cell_type":"code","execution_count":null,"metadata":{"id":"GIjR6Rdw3n2a"},"outputs":[],"source":["import torch\n","import torch.nn as nn\n","import torch.optim as optim\n","import numpy as np\n","import cv2\n","import uuid\n","import os\n","\n","from model import U2NET\n","from torch.autograd import Variable\n","from skimage import io, transform\n","from PIL import Image\n","\n","# Get The Current Directory\n","currentDir = os.path.dirname(__file__)\n","\n","# Functions:\n","# Save Results\n","\n","\n","def save_output(image_name, output_name, pred, d_dir, type):\n","    predict = pred\n","    predict = predict.squeeze()\n","    predict_np = predict.cpu().data.numpy()\n","    im = Image.fromarray(predict_np*255).convert('RGB')\n","    image = io.imread(image_name)\n","    imo = im.resize((image.shape[1], image.shape[0]))\n","    pb_np = np.array(imo)\n","    if type == 'image':\n","        # Make and apply mask\n","        mask = pb_np[:, :, 0]\n","        mask = np.expand_dims(mask, axis=2)\n","        imo = np.concatenate((image, mask), axis=2)\n","        imo = Image.fromarray(imo, 'RGBA')\n","\n","    imo.save(d_dir+output_name)\n","# Remove Background From Image (Generate Mask, and Final Results)\n","\n","\n","def removeBg(imagePath):\n","    inputs_dir = os.path.join(currentDir, 'static/inputs/')\n","    results_dir = os.path.join(currentDir, 'static/results/')\n","    masks_dir = os.path.join(currentDir, 'static/masks/')\n","\n","    # convert string of image data to uint8\n","    with open(imagePath, \"rb\") as image:\n","        f = image.read()\n","        img = bytearray(f)\n","\n","    nparr = np.frombuffer(img, np.uint8)\n","\n","    if len(nparr) == 0:\n","        return '---Empty image---'\n","\n","    # decode image\n","    try:\n","        img = cv2.imdecode(nparr, cv2.IMREAD_COLOR)\n","    except:\n","        # build a response dict to send back to client\n","        return \"---Empty image---\"\n","\n","    # save image to inputs\n","    unique_filename = str(uuid.uuid4())\n","    cv2.imwrite(inputs_dir+unique_filename+'.jpg', img)\n","\n","    # processing\n","    image = transform.resize(img, (320, 320), mode='constant')\n","\n","    tmpImg = np.zeros((image.shape[0], image.shape[1], 3))\n","\n","    tmpImg[:, :, 0] = (image[:, :, 0]-0.485)/0.229\n","    tmpImg[:, :, 1] = (image[:, :, 1]-0.456)/0.224\n","    tmpImg[:, :, 2] = (image[:, :, 2]-0.406)/0.225\n","\n","    tmpImg = tmpImg.transpose((2, 0, 1))\n","    tmpImg = np.expand_dims(tmpImg, 0)\n","    image = torch.from_numpy(tmpImg)\n","\n","    image = image.type(torch.FloatTensor)\n","    image = Variable(image)\n","\n","    d1, d2, d3, d4, d5, d6, d7 = net(image)\n","    pred = d1[:, 0, :, :]\n","    ma = torch.max(pred)\n","    mi = torch.min(pred)\n","    dn = (pred-mi)/(ma-mi)\n","    pred = dn\n","\n","    save_output(inputs_dir+unique_filename+'.jpg', unique_filename +\n","                '.png', pred, results_dir, 'image')\n","    save_output(inputs_dir+unique_filename+'.jpg', unique_filename +\n","                '.png', pred, masks_dir, 'mask')\n","    return \"---Success---\"\n","\n","\n","# ------- Load Trained Model --------\n","print(\"---Loading Model---\")\n","model_name = 'u2net'\n","model_dir = os.path.join(currentDir, 'saved_models',\n","                         model_name, model_name + '.pth')\n","net = U2NET(3, 1)\n","if torch.cuda.is_available():\n","    net.load_state_dict(torch.load(model_dir))\n","    net.cuda()\n","else:\n","    net.load_state_dict(torch.load(model_dir, map_location='cpu'))\n","# ------- Load Trained Model --------\n","\n","\n","print(\"---Removing Background...\")\n","# ------- Call The removeBg Function --------\n","imgPath = \"1.jpg\"  # Change this to your image path\n","print(removeBg(imgPath))\n"]}]}

app/README.md

@@ -0,0 +1,20 @@
+# Image Background Remover Python Script With U2Net
+Remove Backgrounds From Images with u2net Trained model
+
+I was surfing the internet searching for AI tools, and I stumbled upon an Image Background Remover service called Remove.Bg. But I was shocked with the pricing. 0.2$ / Image!
+
+So I decided to take the challenge and build my own Image Background Remover without relying on any APIs or Third Parties.
+
+After Searching for a couple of days, I found a great python project on GitHub, Called U2Net.
+
+A Big Shootout and Thanks to the authors of this project!
+
+I trued my best to make things as simple as possible, so even newbies, can use and run the python script I built.
+
+I created a Short Video Showing The Script & How to use it, and I pointed to two special secrets that may help you build an online income stream with the help of this script. here is the video:
+https://youtu.be/KkhPN7Z4Fy8
+
+Download The Project Source Code With Trainined Model Here:
+https://hasanaboulhasan.com/python-image-background-remover/
+
+Thank You!

app/__init__.py

@@ -0,0 +1,116 @@
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import numpy as np
+import cv2
+import uuid
+import os
+
+from model import U2NET
+from torch.autograd import Variable
+from skimage import io, transform
+from PIL import Image
+
+# Get The Current Directory
+currentDir = os.path.dirname(__file__)
+
+# Functions:
+# Save Results
+
+
+def save_output(image_name, output_name, pred, d_dir, type):
+    predict = pred
+    predict = predict.squeeze()
+    predict_np = predict.cpu().data.numpy()
+    im = Image.fromarray(predict_np*255).convert('RGB')
+    image = io.imread(image_name)
+    imo = im.resize((image.shape[1], image.shape[0]))
+    pb_np = np.array(imo)
+    if type == 'image':
+        # Make and apply mask
+        mask = pb_np[:, :, 0]
+        mask = np.expand_dims(mask, axis=2)
+        imo = np.concatenate((image, mask), axis=2)
+        imo = Image.fromarray(imo, 'RGBA')
+
+    image = imo.convert("RGBA")
+    new_image = Image.new("RGBA", image.size, "WHITE") # Create a white rgba background
+    new_image.paste(image, (0, 0), image)              # Paste the image on the background. Go to the links given below for details.
+    new_image.convert('RGB').save(d_dir+output_name)
+# Remove Background From Image (Generate Mask, and Final Results)
+
+
+def removeBg(imagePath):
+    inputs_dir = os.path.join(currentDir, 'static/inputs/')
+    results_dir = os.path.join(currentDir, 'static/results/')
+    masks_dir = os.path.join(currentDir, 'static/masks/')
+
+    # convert string of image data to uint8
+    with open(imagePath, "rb") as image:
+        f = image.read()
+        img = bytearray(f)
+
+    nparr = np.frombuffer(img, np.uint8)
+
+    if len(nparr) == 0:
+        return '---Empty image---'
+
+    # decode image
+    try:
+        img = cv2.imdecode(nparr, cv2.IMREAD_COLOR)
+    except:
+        # build a response dict to send back to client
+        return "---Empty image---"
+
+    # save image to inputs
+    unique_filename = str(uuid.uuid4())
+    cv2.imwrite(inputs_dir+unique_filename+'.jpg', img)
+
+    # processing
+    image = transform.resize(img, (320, 320), mode='constant')
+
+    tmpImg = np.zeros((image.shape[0], image.shape[1], 3))
+
+    tmpImg[:, :, 0] = (image[:, :, 0]-0.485)/0.229
+    tmpImg[:, :, 1] = (image[:, :, 1]-0.456)/0.224
+    tmpImg[:, :, 2] = (image[:, :, 2]-0.406)/0.225
+
+    tmpImg = tmpImg.transpose((2, 0, 1))
+    tmpImg = np.expand_dims(tmpImg, 0)
+    image = torch.from_numpy(tmpImg)
+
+    image = image.type(torch.FloatTensor)
+    image = Variable(image)
+
+    d1, d2, d3, d4, d5, d6, d7 = net(image)
+    pred = d1[:, 0, :, :]
+    ma = torch.max(pred)
+    mi = torch.min(pred)
+    dn = (pred-mi)/(ma-mi)
+    pred = dn
+
+    save_output(inputs_dir+unique_filename+'.jpg', unique_filename +
+                '.png', pred, results_dir, 'image')
+    save_output(inputs_dir+unique_filename+'.jpg', unique_filename +
+                '.png', pred, masks_dir, 'mask')
+    return "---Success---"
+
+
+# ------- Load Trained Model --------
+print("---Loading Model---")
+model_name = 'u2net'
+model_dir = os.path.join(currentDir, 'saved_models',
+                         model_name, model_name + '.pth')
+net = U2NET(3, 1)
+if torch.cuda.is_available():
+    net.load_state_dict(torch.load(model_dir))
+    net.cuda()
+else:
+    net.load_state_dict(torch.load(model_dir, map_location='cpu'))
+# ------- Load Trained Model --------
+
+
+print("---Removing Background...")
+# ------- Call The removeBg Function --------
+imgPath = "1.jpg"  # Change this to your image path
+print(removeBg(imgPath))

app/config.py

@@ -0,0 +1,8 @@
+import json
+
+with open('/etc/config.json') as config_file:
+	config = json.load(config_file)
+
+class Config:
+	SECRET_KEY = config.get('SECRET_KEY')
+