241113_main_weapons_dataset_make.ipynb

{"cells":[{"cell_type":"code","execution_count":1,"metadata":{"executionInfo":{"elapsed":422,"status":"ok","timestamp":1675493425499,"user":{"displayName":"Naoki Kimura","userId":"07848497433963924783"},"user_tz":-540},"id":"PhyNj6nTzj6m"},"outputs":[],"source":["import glob\n","import re\n","import os\n","from PIL import Image,ImageOps,ImageFilter\n","import matplotlib.pyplot as plt\n","import numpy as np\n","import random\n","import time\n","import unicodedata\n","\n","# -*- coding: utf-8 -*-\n","\n","yolo_size = 416\n","\n","resize_size = (200, 200)\n","\n"]},{"cell_type":"code","execution_count":2,"metadata":{},"outputs":[{"name":"stdout","output_type":"stream","text":["Files already downloaded and verified\n","背景画像の数: 50000\n"]}],"source":["# CIFAR10をバックグラウンド画像として使用\n","from torchvision.datasets import CIFAR10\n","import torchvision.transforms as transforms\n","import numpy as np\n","\n","# データセットの取得\n","dataset = CIFAR10(root='./data', train=True, download=True)\n","\n","# NumPy配列からPIL Imageに変換\n","base_images = []\n","for img, _ in dataset:\n","    # CIFAR10の画像はすでにPIL Imageとして取得されるので、\n","    # そのまま使用可能\n","    base_images.append(img)\n","\n","base_images_length = len(base_images)\n","print(f\"背景画像の数: {base_images_length}\")"]},{"cell_type":"code","execution_count":3,"metadata":{},"outputs":[{"name":"stdout","output_type":"stream","text":["['E-liter-4K', 'Nautilus-47', 'Dapple-Dualies-Nouveau', 'Custom_Explosher', 'Order_Blaster_Replica', 'Octobrush', 'Flingza-Roller', 'Luna-Blaster', 'Annaki_Splattershot_Nova', 'Splat-Dualies', 'Tentatek-Splattershot', 'Neo_Sploosh-o-matic', 'Splattershot', 'Order_Roller_Replica', 'Custom_Douser_Dualies_FF', 'Enperry_Splat_Dualies', 'Dark-Tetra-Dualies', 'Explosher', 'Custom_Hydra_Splatling', 'Slosher', 'Mini-Splatling', 'Custom_Goo_Tuber', 'Tenta-Brella', 'Glooga-Dualies', 'Inkbrush-Nouveau', 'Splash-o-matic', 'Mint_Decavitator', 'Dapple-Dualies', 'Slosher-Deco', 'Goo-Tuber', 'Undercover-Brella', 'Order_Splatling_Replica', 'Splat-Charger', 'Order_Shot_Replica', 'Gold_Dynamo_Roller', '96-Gal', 'L-3_Nozzlenose_D', 'Ballpoint_Splatling_Nouveau', 'Jet-Squelcher', 'Bloblobber_Deco', 'LACT-450', 'Carbon-Roller-Deco', 'Aerospray-RG', 'Zink-Mini-Splatling', 'Range-Blaster', 'Blaster', 'Hero-Shooter-Replica', 'Custom_Range_Blaster', 'Splat-Brella', 'Order_Stringer_Replica', 'Inkbrush', 'Inkline_Tri-Stringer', 'Clash_Blaster_Neo', 'N-ZAP85', 'Light_Tetra_Dualies', 'L-3-Nozzlenose', 'Custom_Jet_Squelcher', 'Heavy-Splatling', 'Splatana-Stamper', 'Splattershot-Pro', 'Splattershot-Nova', 'Rapid-Blaster-Pro', 'Sploosh-o-matic', 'Neo_Splash-o-matic', 'Custom_E-liter_4K_Scope', 'Foil_Flingza_Roller', 'Heavy_Edit_Splatling_Nouveau', 'Tri-Stringer', 'Splattershot-Jr', 'Hydra-Splatling', 'Heavy_Splatling_Deco', 'Order_Brella_Replica', 'Nautilus_79', 'Order_Slosher_Replica', 'Glooga_Dualies_Deco', 'H-3-Nozzlenose', 'Big_Swig_Roller_Express', 'Foil_Squeezer', 'Order_Splatana_Replica', 'Big-Swig-Roller', 'Heavy_Edit_Splatling', 'Custom_E-liter_4K', 'Dynamo-Roller', 'Forge_Splattershot_Pro', 'Charcoal_Decavitator', 'Dualie-Squelchers', 'Sloshing-Machine', 'Bloblobber', 'Douser_Dualies_FF', 'Custom_Wellstring_V', 'Tri-Slosher', 'Splatana-Wiper', 'Custom-Splattershot-Jr', 'Aerospray', 'Clash-Blaster', 'Ballpoint-Splatling', 'Classic_Squiffer', 'Krak-On_Splat_Roller', 'Order_Dualie_Replicas', 'Carbon-Roller', 'Dread_Wringer_D', 'Luna-Blaster-Neo', 'Squeezer', 'Splat-Roller', 'Rapid-Blaster', 'Dread_Wringer', 'Custom_Blaster', 'Order_Charger_Replica', 'Snipewriter-5H', 'Squiffer', 'New_Squiffer', 'Custom_Dualie_Squelchers', 'Bamboozler_14_Mk_II', 'H-3_Nozzlenose_D', '52-Gal', 'Bamboozler-14-Mk-I']\n"]}],"source":["icons = glob.glob(\"main_icons/*.png\")\n","\n","class_list = []\n","for path in icons:\n","    # main_icons/Weapon-Name.png の形式から武器名だけを抽出\n","    weapon_name = path.replace('main_icons/', '').replace('.png', '')\n","    class_list.append(weapon_name)\n","print(class_list)\n","\n","f = open('weapon_list.txt', 'w')\n","for x in class_list:\n","    f.write(str(x) + \"\\n\")\n","f.close()"]},{"cell_type":"code","execution_count":4,"metadata":{"executionInfo":{"elapsed":248,"status":"ok","timestamp":1675493618244,"user":{"displayName":"Naoki Kimura","userId":"07848497433963924783"},"user_tz":-540},"id":"fm2Yd_lxzj6p"},"outputs":[],"source":["import cv2\n","from PIL import ImageEnhance    #ImageEnhanceのインポート\n","def take_path_generate_dataset(weapons_dir_path):\n","    label_info = []\n","    # base_pathはすでにPIL Imageオブジェクト\n","    base = base_images[random.randint(0, base_images_length-1)]\n","    base = base.resize(resize_size)  # リサイズ\n","    base = base.filter(ImageFilter.BLUR)\n","    \n","    for i in range(0,1):\n","        logo_relative_positions = [random.uniform(-0.5, 0.0), random.uniform(-0.5, 0.0)]\n","        logo_position = (int(100*logo_relative_positions[0]), int(100*logo_relative_positions[1]))\n","\n","        label = random.choice(class_list)\n","        \n","        logo = Image.open(f'{weapons_dir_path}/{label}.png')\n","        logo_w, logo_h = logo.size\n","        \n","        logo_scale_x = random.uniform(0.7, 1.1)\n","        logo_scale_y = random.uniform(0.7, 1.1)\n","\n","        logo = logo.resize((int(logo_w * logo_scale_x), int(logo_h * logo_scale_y)))\n","        \n","        #色のオーグメンテーション\n","        logo = ImageEnhance.Color(logo)\n","        logo = logo.enhance(random.uniform(0, 1))\n","        \n","        if random.randint(0,10) < 2:\n","            logo = logo.rotate(random.randint(-10,10), expand=True)\n","        \n","        base.paste(logo, logo_position, logo)\n","        \n","    return label, base\n"]},{"cell_type":"code","execution_count":5,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":286},"executionInfo":{"elapsed":892,"status":"ok","timestamp":1675493625373,"user":{"displayName":"Naoki Kimura","userId":"07848497433963924783"},"user_tz":-540},"id":"FVHT3ujxzj6p","outputId":"820e3f23-df13-4646-d830-ae75e4b95603"},"outputs":[{"name":"stdout","output_type":"stream","text":["Heavy_Edit_Splatling\n"]},{"data":{"image/png":"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","text/plain":["<Figure size 640x480 with 1 Axes>"]},"metadata":{},"output_type":"display_data"}],"source":["label, base = take_path_generate_dataset(\"main_icons\")\n","print(label)\n","randomized_size = random.randint(32,64)\n","base = base.resize((randomized_size,randomized_size))\n","base = base.resize((64,64))\n","plt.imshow(np.array(base))\n","plt.show()"]},{"cell_type":"code","execution_count":6,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":376385,"status":"ok","timestamp":1675494067695,"user":{"displayName":"Naoki Kimura","userId":"07848497433963924783"},"user_tz":-540},"id":"EcRAQitBzj6p","outputId":"4869d17c-8d77-4eb9-efa0-95ea64e2fff8"},"outputs":[{"name":"stdout","output_type":"stream","text":["main_training_dataset/Explosher\n","main_training_dataset/Douser_Dualies_FF\n","main_training_dataset/Zink-Mini-Splatling\n","main_training_dataset/Aerospray-RG\n","main_training_dataset/Ballpoint-Splatling\n","main_training_dataset/Ballpoint_Splatling_Nouveau\n","main_training_dataset/Carbon-Roller-Deco\n","main_training_dataset/Luna-Blaster-Neo\n","main_training_dataset/Splat-Brella\n","main_training_dataset/H-3-Nozzlenose\n","main_training_dataset/Aerospray\n","main_training_dataset/Gold_Dynamo_Roller\n","main_training_dataset/Rapid-Blaster\n","main_training_dataset/Splatana-Stamper\n","main_training_dataset/Heavy_Splatling_Deco\n","main_training_dataset/LACT-450\n","main_training_dataset/Custom_Jet_Squelcher\n","main_training_dataset/Heavy-Splatling\n","main_training_dataset/Splat-Charger\n","main_training_dataset/Big_Swig_Roller_Express\n","main_training_dataset/Custom_Range_Blaster\n","main_training_dataset/Jet-Squelcher\n","main_training_dataset/Dark-Tetra-Dualies\n","main_training_dataset/Splattershot-Pro\n","main_training_dataset/Order_Splatana_Replica\n","main_training_dataset/Dynamo-Roller\n","main_training_dataset/Custom_Hydra_Splatling\n","main_training_dataset/Sloshing-Machine\n","main_training_dataset/Custom_Blaster\n","main_training_dataset/Slosher-Deco\n","main_training_dataset/Inkbrush-Nouveau\n","main_training_dataset/Splattershot\n","main_training_dataset/Bamboozler-14-Mk-I\n","main_training_dataset/Undercover-Brella\n","main_training_dataset/Slosher\n","main_training_dataset/Order_Shot_Replica\n","main_training_dataset/Nautilus_79\n","main_training_dataset/Mini-Splatling\n","main_training_dataset/Inkbrush\n","main_training_dataset/Snipewriter-5H\n","main_training_dataset/Tri-Stringer\n","main_training_dataset/Carbon-Roller\n","main_training_dataset/Order_Slosher_Replica\n","main_training_dataset/Light_Tetra_Dualies\n","main_training_dataset/Hero-Shooter-Replica\n","main_training_dataset/E-liter-4K\n","main_training_dataset/Tenta-Brella\n","main_training_dataset/Clash-Blaster\n","main_training_dataset/Custom_Dualie_Squelchers\n","main_training_dataset/New_Squiffer\n","main_training_dataset/Dread_Wringer\n","main_training_dataset/Nautilus-47\n","main_training_dataset/Custom_Goo_Tuber\n","main_training_dataset/Enperry_Splat_Dualies\n","main_training_dataset/Heavy_Edit_Splatling_Nouveau\n","main_training_dataset/Custom_Explosher\n","main_training_dataset/Octobrush\n","main_training_dataset/Range-Blaster\n","main_training_dataset/Luna-Blaster\n","main_training_dataset/Neo_Splash-o-matic\n","main_training_dataset/Neo_Sploosh-o-matic\n","main_training_dataset/Rapid-Blaster-Pro\n","main_training_dataset/H-3_Nozzlenose_D\n","main_training_dataset/Splattershot-Nova\n","main_training_dataset/Order_Charger_Replica\n","main_training_dataset/Splat-Roller\n","main_training_dataset/Tentatek-Splattershot\n","main_training_dataset/52-Gal\n","main_training_dataset/Big-Swig-Roller\n","main_training_dataset/Splash-o-matic\n","main_training_dataset/Goo-Tuber\n","main_training_dataset/Glooga_Dualies_Deco\n","main_training_dataset/Foil_Squeezer\n","main_training_dataset/Mint_Decavitator\n","main_training_dataset/Foil_Flingza_Roller\n","main_training_dataset/Order_Blaster_Replica\n","main_training_dataset/Bamboozler_14_Mk_II\n","main_training_dataset/Glooga-Dualies\n","main_training_dataset/Charcoal_Decavitator\n","main_training_dataset/Custom_Wellstring_V\n","main_training_dataset/Forge_Splattershot_Pro\n","main_training_dataset/Krak-On_Splat_Roller\n","main_training_dataset/Order_Brella_Replica\n","main_training_dataset/96-Gal\n","main_training_dataset/Custom-Splattershot-Jr\n","main_training_dataset/Dualie-Squelchers\n","main_training_dataset/Custom_E-liter_4K_Scope\n","main_training_dataset/Inkline_Tri-Stringer\n","main_training_dataset/Hydra-Splatling\n","main_training_dataset/Squeezer\n","main_training_dataset/Custom_E-liter_4K\n","main_training_dataset/Tri-Slosher\n","main_training_dataset/Splat-Dualies\n","main_training_dataset/L-3_Nozzlenose_D\n","main_training_dataset/Splatana-Wiper\n","main_training_dataset/Clash_Blaster_Neo\n","main_training_dataset/Dapple-Dualies\n","main_training_dataset/Custom_Douser_Dualies_FF\n","main_training_dataset/Dread_Wringer_D\n","main_training_dataset/Blaster\n","main_training_dataset/Order_Splatling_Replica\n","main_training_dataset/L-3-Nozzlenose\n","main_training_dataset/Order_Stringer_Replica\n","main_training_dataset/Splattershot-Jr\n","main_training_dataset/Bloblobber\n","main_training_dataset/Dapple-Dualies-Nouveau\n","main_training_dataset/Heavy_Edit_Splatling\n","main_training_dataset/Bloblobber_Deco\n","main_training_dataset/Annaki_Splattershot_Nova\n","main_training_dataset/Squiffer\n","main_training_dataset/Order_Roller_Replica\n","main_training_dataset/Flingza-Roller\n","main_training_dataset/N-ZAP85\n","main_training_dataset/Order_Dualie_Replicas\n","main_training_dataset/Classic_Squiffer\n","main_training_dataset/Sploosh-o-matic\n"]}],"source":["from PIL import ImageFile\n","ImageFile.LOAD_TRUNCATED_IMAGES = True\n","\n","dataset_directory = \"main_training_dataset\"\n","!rm -rf main_training_dataset\n","import os\n","\n","if not os.path.exists(dataset_directory):\n","   os.makedirs(dataset_directory)\n","\n","for i in range(10000):\n","        label, base = take_path_generate_dataset(\"main_icons\")\n","        if not os.path.exists(f\"{dataset_directory}/{label}\"):\n","            print(f\"{dataset_directory}/{label}\")\n","            os.mkdir(f\"{dataset_directory}/{label}\")\n","        randomized_size = random.randint(32,64)\n","        base = base.resize((randomized_size,randomized_size))\n","        base = base.resize((64,64))\n","        base.save(f\"{dataset_directory}/{label}/{time.time()}.jpg\", quality=95)"]},{"cell_type":"code","execution_count":null,"metadata":{},"outputs":[],"source":[]}],"metadata":{"colab":{"provenance":[{"file_id":"1ZIFQWW2l4KhfJE4vb4BcTpmMomYayqNN","timestamp":1675504264433}]},"kernelspec":{"display_name":"base","language":"python","name":"python3"},"language_info":{"codemirror_mode":{"name":"ipython","version":3},"file_extension":".py","mimetype":"text/x-python","name":"python","nbconvert_exporter":"python","pygments_lexer":"ipython3","version":"3.10.14"},"orig_nbformat":4,"vscode":{"interpreter":{"hash":"d4d1e4263499bec80672ea0156c357c1ee493ec2b1c70f0acce89fc37c4a6abe"}}},"nbformat":4,"nbformat_minor":0}