{"id":13325,"date":"2025-05-01T09:12:03","date_gmt":"2025-05-01T00:12:03","guid":{"rendered":"https:\/\/blogs.mathworks.com\/japan-community\/?p=13325"},"modified":"2025-05-06T14:50:06","modified_gmt":"2025-05-06T05:50:06","slug":"llm-enhanced-anomaly-classification-for-images-jp","status":"publish","type":"post","link":"https:\/\/blogs.mathworks.com\/japan-community\/2025\/05\/01\/llm-enhanced-anomaly-classification-for-images-jp\/","title":{"rendered":"LLM\u3067\u9032\u5316\u3059\u308b\u753b\u50cf\u306e\u7570\u5e38\u5206\u985e\uff1aAI\u306b\u3088\u308b\u81ea\u52d5\u89e3\u6790\u3068\u81ea\u7136\u8a00\u8a9e\u8aac\u660e"},"content":{"rendered":"<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u203b\u3053\u306e\u6295\u7a3f\u306f 2025 \u5e74 4 \u6708 21 \u65e5\u306b <a href=\"https:\/\/blogs.mathworks.com\/deep-learning\/2025\/04\/21\/llm-enhanced-anomaly-classification-for-images\/?from=jp\">Artificial Intelligence \u3078 \u6295\u7a3f<\/a>\u3055\u308c\u305f\u3082\u306e\u306e\u6284\u8a33\u3067\u3059\u3002<\/div>\n<hr \/>\n<h6><\/h6>\n<p>\u753b\u50cf\u5185\u306e\u4e88\u671f\u3057\u306a\u3044\u30d1\u30bf\u30fc\u30f3\u306b\u3001\u610f\u5473\u306e\u3042\u308b\u30ab\u30c6\u30b4\u30ea\u3084\u8aac\u660e\u3092\u4ed8\u3051\u308b\u3053\u3068\uff08= \u753b\u50cf\u306e\u7570\u5e38\u5206\u985e\uff09\u306f\u3001\u54c1\u8cea\u7ba1\u7406\u306e\u81ea\u52d5\u5316\u3084\u533b\u7642\u8a3a\u65ad\u306e\u9ad8\u5ea6\u5316\u306b\u5f79\u7acb\u3061\u307e\u3059\u3002\u672c\u30d6\u30ed\u30b0\u3067\u306f\u3001\u753b\u50cf\u30c7\u30fc\u30bf\u5411\u3051\u306e\u7570\u5e38\u5206\u985e\u30e2\u30c7\u30eb\u3067\u3042\u308b <strong>EfficientAD<\/strong> \u3068\u3001<strong>\u5927\u898f\u6a21\u8a00\u8a9e\u30e2\u30c7\u30eb\uff08LLM\uff09<\/strong>\u3092\u6d3b\u7528\u3057\u3001MATLAB \u3067\u7570\u5e38\u5206\u985e\u3092\u884c\u3046\u65b9\u6cd5\u3092\u7d39\u4ecb\u3057\u307e\u3059\u3002<\/p>\n<h6><\/h6>\n<p>\u3053\u3053\u3067\u3054\u7d39\u4ecb\u3059\u308b\u30ef\u30fc\u30af\u30d5\u30ed\u30fc\u306f\u3001\u753b\u50cf\u30c7\u30fc\u30bf\u306b\u5bfe\u3059\u308b\u7570\u5e38\u5206\u985e\u3092\u81ea\u52d5\u5316\u3059\u308b\u3060\u3051\u3067\u306a\u304f\u3001LLM \u306e\u52a9\u3051\u3092\u501f\u308a\u3066\u3001\u5206\u985e\u7d50\u679c\u306b\u3064\u3044\u3066\u81ea\u7136\u8a00\u8a9e\u3067\u306e\u8aac\u660e\u3082\u751f\u6210\u3057\u307e\u3059\u3002\u307e\u305f\u3001LLM \u3092\u6d3b\u7528\u3057\u305f\u7570\u5e38\u5206\u985e\u3092\u884c\u3046\u305f\u3081\u306b\u5fc5\u8981\u306a\u30b3\u30fc\u30c9\u306f <a href=\"https:\/\/github.com\/matlab-deep-learning\/zero-shot-anomaly-classification-with-EfficientAD-and-LLM\" class=\"external\" rel=\"nofollow\" target=\"_blank\">GitHub\u30ea\u30dd\u30b8\u30c8\u30ea<\/a>\u306b\u3042\u308a\u307e\u3059\u304c\u3001\u3053\u3053\u3067\u3082\u4e00\u90e8\u3054\u7d39\u4ecb\u3057\u307e\u3059\u3002<\/p>\n<h6><\/h6>\n<p><img decoding=\"async\" loading=\"lazy\" class=\"alignnone wp-image-17282 size-full\" src=\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2025\/04\/anomaly_classification_intro.png\" alt=\"Showing concept of anomaly classification with LLM-generated explanations for fabric images\" width=\"1809\" height=\"874\" \/><\/p>\n<h6><\/h6>\n<p>&nbsp;<\/p>\n<h6><\/h6>\n<p style=\"font-size: 20px; color: #c04c0b;\"><strong>\u91cd\u8981\u306a\u30b3\u30f3\u30bb\u30d7\u30c8<\/strong><\/p>\n<p style=\"font-size: 18px;\"><strong>\u7570\u5e38\u5206\u985e\u3068\u306f\uff1f<\/strong><\/p>\n<p class=\"\" data-start=\"13\" data-end=\"160\">\u7570\u5e38\u5206\u985e\u3068\u306f\u3001\u753b\u50cf\u30c7\u30fc\u30bf\u306b\u304a\u3051\u308b\u901a\u5e38\u3068\u306f\u7570\u306a\u308b\u30d1\u30bf\u30fc\u30f3\u306b\u5bfe\u3057\u3066\u3001\u610f\u5473\u306e\u3042\u308b\u30e9\u30d9\u30eb\u3084\u30ab\u30c6\u30b4\u30ea\u3092\u5272\u308a\u5f53\u3066\u308b\u4f5c\u696d\u3067\u3059\u3002\u4e00\u822c\u7684\u306a\u753b\u50cf\u5206\u985e\u304c\u300c\u753b\u50cf\u306b\u4f55\u304c\u5199\u3063\u3066\u3044\u308b\u304b\u300d\u3092\u8b58\u5225\u3059\u308b\u306e\u306b\u5bfe\u3057\u3001\u7570\u5e38\u5206\u985e\u306f\u300c\u3069\u306e\u3088\u3046\u306a\u7570\u5e38\u304b\u300d\u306b\u7279\u5316\u3057\u3066\u304a\u308a\u3001\u4e0d\u898f\u5247\u306a\u30d1\u30bf\u30fc\u30f3\u3092\u691c\u51fa\u3057\u3001\u305d\u308c\u304c\u3069\u306e\u7a2e\u985e\u306e\u7570\u5e38\u306b\u8a72\u5f53\u3059\u308b\u304b\u3092\u5224\u65ad\u3057\u307e\u3059\u3002\u305f\u3068\u3048\u3070\u88fd\u9020\u696d\u3067\u306e\u7570\u5e38\u5206\u985e\u3068\u3057\u3066\u306f\u88fd\u54c1\u8868\u9762\u306e\u50b7\u3001\u3078\u3053\u307f\u3001\u5857\u88c5\u4e0d\u826f\u306a\u3069\u3092\u533a\u5225\u3059\u308b\u3053\u3068\u304c\u6c42\u3081\u3089\u308c\u308b\u304b\u3082\u3057\u308c\u307e\u305b\u3093\u3002\u307e\u305f\u533b\u7528\u753b\u50cf\u306e\u5834\u5408\u306f\u3001\u75c5\u5909\u304c\u826f\u6027\u304b\u60aa\u6027\u304b\u306e\u5206\u985e\u304c\u8003\u3048\u3089\u308c\u307e\u3059\u3002<\/p>\n<h6><\/h6>\n<p class=\"\" data-start=\"245\" data-end=\"397\">\u7570\u5e38\u5206\u985e\u306e\u96e3\u3057\u3055\u3068\u3057\u3066\u306f\u3001\u7570\u5e38\u304c\u5f80\u3005\u306b\u3057\u3066\u7a00\u304b\u3064\u591a\u69d8\u3067\u3042\u308a\u3001\u6587\u8108\u306b\u3082\u4f9d\u5b58\u3059\u308b\u3068\u3044\u3046\u70b9\u304c\u6319\u3052\u3089\u308c\u307e\u3059\u3002\u3053\u308c\u306b\u3088\u308a\u7570\u5e38\u5206\u985e\u306f\u901a\u5e38\u306e\u5206\u985e\u30bf\u30b9\u30af\u3088\u308a\u3082\u8907\u96d1\u306b\u306a\u308a\u307e\u3059\u3002\u52b9\u679c\u7684\u306a\u7570\u5e38\u5206\u985e\u30e2\u30c7\u30eb\u306f\u3001\u5358\u306b\u7570\u5e38\u3092\u691c\u51fa\u3059\u308b\u3060\u3051\u3067\u306a\u304f\u3001\u305d\u306e\u6027\u8cea\u30fb\u6df1\u523b\u3055\u3092\u89e3\u91c8\u3057\u3001\u5834\u5408\u306b\u3088\u3063\u3066\u306f\u9069\u5207\u306a\u5bfe\u5fdc\u3092\u63d0\u6848\u3059\u308b\u5fc5\u8981\u304c\u3042\u308a\u307e\u3059\u3002\u305d\u306e\u305f\u3081\u306b\u306f\u3001\u30e2\u30c7\u30eb\u306b\u9ad8\u3044\u7cbe\u5ea6\u3060\u3051\u3067\u306a\u304f\u3001\u8aac\u660e\u53ef\u80fd\u6027\u3084\u9069\u5fdc\u6027\u3082\u6c42\u3081\u3089\u308c\u307e\u3059\u3002\u7279\u306b\u3001\u65b0\u3057\u3044\u7570\u5e38\u3084\u3053\u308c\u307e\u3067\u306b\u898b\u305f\u3053\u3068\u306e\u306a\u3044\u7570\u5e38\u304c\u767b\u5834\u3057\u5f97\u308b\u74b0\u5883\u3067\u306f\u3001\u3053\u308c\u3089\u306e\u80fd\u529b\u304c\u4e0d\u53ef\u6b20\u3067\u3059\u3002<\/p>\n<h6><\/h6>\n<p style=\"font-size: 18px;\"><strong>EfficientAD\uff1a\u30ea\u30a2\u30eb\u30bf\u30a4\u30e0\u7570\u5e38\u5206\u985e\u306e\u305f\u3081\u306e\u30e2\u30c7\u30eb<\/strong><\/p>\n<p class=\"\" data-start=\"35\" data-end=\"219\">EfficientAD \u306f\u30ea\u30a2\u30eb\u30bf\u30a4\u30e0\u306a\u7523\u696d\u7528\u9014\u3092\u60f3\u5b9a\u3057\u3066\u8a2d\u8a08\u3055\u308c\u305f\u3001\u9ad8\u6027\u80fd\u306a\u5916\u89b3\u7570\u5e38\u5206\u985e\u30e2\u30c7\u30eb\u3067\u3059\u3002\u3053\u306e\u30e2\u30c7\u30eb\u306f\u300c\u30b9\u30c1\u30e5\u30fc\u30c7\u30f3\u30c8\u30fb\u30c6\u30a3\u30fc\u30c1\u30e3\u30fc\u30a2\u30fc\u30ad\u30c6\u30af\u30c1\u30e3\u300d\u3092\u63a1\u7528\u3057\u3066\u304a\u308a\u3001\u8efd\u91cf\u306a\u30b9\u30c1\u30e5\u30fc\u30c7\u30f3\u30c8\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u304c\u3001\u4e8b\u524d\u5b66\u7fd2\u3055\u308c\u305f\u5927\u898f\u6a21\u306a\u30c6\u30a3\u30fc\u30c1\u30e3\u30fc\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u306e\u7279\u5fb4\u8868\u73fe\u3092\u6a21\u5023\u3059\u308b\u3088\u3046\u306b\u5b66\u7fd2\u3055\u308c\u307e\u3059\u3002\u5b66\u7fd2\u306b\u306f\u300c\u6b63\u5e38\u300d\u307e\u305f\u306f\u300c\u671f\u5f85\u3055\u308c\u308b\u72b6\u614b\u300d\u3092\u8868\u3059\u753b\u50cf\u304c\u4f7f\u7528\u3055\u308c\u307e\u3059\u3002\u63a8\u8ad6\u6642\u306b\u306f\u3001\u30b9\u30c1\u30e5\u30fc\u30c7\u30f3\u30c8\u3068\u30c6\u30a3\u30fc\u30c1\u30e3\u30fc\u306e\u51fa\u529b\u306e\u5dee\u7570\u3092\u3082\u3068\u306b\u3001\u7570\u5e38\u30d1\u30bf\u30fc\u30f3\u3092\u5206\u985e\u3057\u307e\u3059\u3002\u3053\u306e\u30a2\u30fc\u30ad\u30c6\u30af\u30c1\u30e3\u306f\u9ad8\u901f\u304b\u3064\u9ad8\u7cbe\u5ea6\u3067\u3042\u308a\u30011 \u679a\u306e\u753b\u50cf\u3092 2 \u30df\u30ea\u79d2\u672a\u6e80\u3067\u51e6\u7406\u3067\u304d\u3001\u6700\u65b0\u306e GPU \u4e0a\u3067\u306f\u6bce\u79d2 600 \u679a\u4ee5\u4e0a\u306e\u753b\u50cf\u51e6\u7406\u304c\u53ef\u80fd\u3067\u3059\u3002<\/p>\n<h6><\/h6>\n<p class=\"\" data-start=\"342\" data-end=\"390\">EfficientAD \u306e\u5927\u304d\u306a\u7279\u9577\u306f\u3001\u300c\u69cb\u9020\u7684\u7570\u5e38\u300d\u3068\u300c\u8ad6\u7406\u7684\u7570\u5e38\u300d\u306e\u4e21\u65b9\u306b\u5bfe\u5fdc\u3067\u304d\u308b\u70b9\u3067\u3059\u3002<\/p>\n<ul data-start=\"392\" data-end=\"479\">\n<li class=\"\" data-start=\"392\" data-end=\"434\">\n<p class=\"\" data-start=\"394\" data-end=\"434\"><strong data-start=\"394\" data-end=\"403\">\u69cb\u9020\u7684\u7570\u5e38<\/strong>\uff1a\u8868\u9762\u306e\u50b7\u3084\u305a\u308c\u306a\u3069\u3001\u5c40\u6240\u7684\u306a\u7279\u5fb4\u306e\u30ba\u30ec\u306b\u3088\u3063\u3066\u691c\u51fa\u3055\u308c\u307e\u3059\u3002<\/p>\n<\/li>\n<li class=\"\" data-start=\"435\" data-end=\"479\">\n<p class=\"\" data-start=\"437\" data-end=\"479\"><strong data-start=\"437\" data-end=\"446\">\u8ad6\u7406\u7684\u7570\u5e38<\/strong>\uff1a\u610f\u5473\u7684\u306a\u69cb\u9020\u306e\u77db\u76fe\u3001\u4f8b\u3048\u3070\u4e88\u671f\u3057\u306a\u3044\u7269\u4f53\u306e\u914d\u7f6e\u306a\u3069\u304c\u8a72\u5f53\u3057\u307e\u3059\u3002<\/p>\n<\/li>\n<\/ul>\n<p class=\"\" data-start=\"481\" data-end=\"594\">\u3053\u306e\u3088\u3046\u306a\u9ad8\u5ea6\u306a\u5206\u985e\u3092\u53ef\u80fd\u306b\u3059\u308b\u305f\u3081\u306b\u3001EfficientAD \u306f\u30aa\u30fc\u30c8\u30a8\u30f3\u30b3\u30fc\u30c0\u30fc\u30e2\u30b8\u30e5\u30fc\u30eb\u3092\u7d44\u307f\u8fbc\u307f\u3001\u753b\u50cf\u5168\u4f53\u306e\u610f\u5473\u7684\u306a\u69cb\u9020\u3092\u6349\u3048\u308b\u3053\u3068\u304c\u3067\u304d\u308b\u3088\u3046\u306b\u3057\u3066\u3044\u307e\u3059\u3002\u3053\u308c\u306b\u3088\u308a\u3001\u5358\u7d14\u306a\u898b\u305f\u76ee\u306e\u9055\u3044\u3092\u8d85\u3048\u305f\u5206\u985e\u304c\u53ef\u80fd\u306b\u306a\u308a\u307e\u3059\u3002\u305d\u306e\u7d50\u679c\u3001EfficientAD \u306f\u8a08\u7b97\u8cc7\u6e90\u304c\u9650\u3089\u308c\u305f\u74b0\u5883\u3067\u3082\u5b89\u5b9a\u3057\u305f\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u3092\u767a\u63ee\u3067\u304d\u308b\u3001\u8fc5\u901f\u3067\u67d4\u8edf\u306a\u30e2\u30c7\u30eb\u3068\u306a\u3063\u3066\u304a\u308a\u3001\u73fe\u5b9f\u306e\u30a2\u30d7\u30ea\u30b1\u30fc\u30b7\u30e7\u30f3\u306b\u975e\u5e38\u306b\u9069\u3057\u305f\u3082\u306e\u3068\u306a\u3063\u3066\u3044\u307e\u3059\u3002<\/p>\n<h6><\/h6>\n<p>R2024b \u4ee5\u964d\u3001EfficientAD \u306f MATLAB \u3067 <a href=\"https:\/\/www.mathworks.com\/help\/vision\/ref\/efficientadanomalydetector.html\">efficientADAnomalyDetector<\/a> \u30aa\u30d6\u30b8\u30a7\u30af\u30c8\u3068\u3057\u3066\u5229\u7528\u53ef\u80fd\u306b\u306a\u308a\u307e\u3057\u305f\u3002\u3053\u306e EfficientAD \u306f\u3001<a href=\"https:\/\/www.mathworks.com\/matlabcentral\/fileexchange\/116555-autom%C3%A5ated-visual-inspection-library-for-computer-vision-toolbox\">Automated Visual Inspection Library for Computer Vision Toolbox<\/a> \u306b\u542b\u307e\u308c\u3066\u304a\u308a\u3001\u7570\u5e38\u5206\u985e\u3001\u7570\u5e38\u691c\u51fa\u3001\u7269\u4f53\u691c\u51fa\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u306e\u5b66\u7fd2\u30fb\u8a55\u4fa1\u30fb\u30c7\u30d7\u30ed\u30a4\u306e\u305f\u3081\u306e\u95a2\u6570\u7fa4\u3092\u63d0\u4f9b\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n<h6><\/h6>\n<p style=\"font-size: 18px;\"><strong>LLM \u306b\u3088\u308b\u7570\u5e38\u5206\u985e\u306e\u9ad8\u5ea6\u5316<\/strong><\/p>\n<p class=\"\" data-start=\"21\" data-end=\"144\">EfficientAD \u304c\u8996\u899a\u7684\u306a\u7570\u5e38\u3092\u5206\u985e\u3059\u308b\u4e00\u65b9\u3067\u3001\u5927\u898f\u6a21\u8a00\u8a9e\u30e2\u30c7\u30eb\uff08LLM\uff09 \u306f\u305d\u306e\u7d50\u679c\u306b\u6587\u8108\u3092\u4e0e\u3048\u3001\u5c02\u9580\u9818\u57df\u306b\u5fdc\u3058\u305f\u8aac\u660e\u3092\u63d0\u4f9b\u3057\u3001<a href=\"https:\/\/www.mathworks.com\/discovery\/interpretability.html\">\u7d50\u679c\u306e\u8aac\u660e\u6027<\/a>\u3092\u9ad8\u3081\u308b\u3053\u3068\u304c\u3067\u304d\u307e\u3059\u3002\u3053\u306e\u3088\u3046\u306a\u6587\u8108\u306b\u304a\u3044\u3066\u3001LLM \u306f\u4ee5\u4e0b\u306e\u3088\u3046\u306a\u5f79\u5272\u3092\u679c\u305f\u3057\u307e\u3059\uff1a<\/p>\n<ul data-start=\"146\" data-end=\"229\">\n<li class=\"\" data-start=\"146\" data-end=\"174\">\u7570\u5e38\u30de\u30c3\u30d7\u3092\u4eba\u9593\u304c\u7406\u89e3\u3057\u3084\u3059\u3044\u5f62\u3067\u89e3\u91c8\u30fb\u8868\u73fe\u3059\u308b<\/li>\n<li>\u904e\u53bb\u306e\u7570\u5e38\u8a18\u9332\u306b\u57fa\u3065\u304d\u3001\u6f5c\u5728\u7684\u306a\u539f\u56e0\u3092\u63a8\u6e2c\u3059\u308b<\/li>\n<li>\u81ea\u52d5\u7570\u5e38\u30ec\u30dd\u30fc\u30c8\u4f5c\u6210\u3084\u5bfe\u5fdc\u7b56\u306e\u63d0\u6848\u3092\u652f\u63f4\u3059\u308b<\/li>\n<\/ul>\n<p>MATLAB \u3067\u306f\u3001API \u7d4c\u7531\u307e\u305f\u306f\u30ed\u30fc\u30ab\u30eb\u306b\u30e2\u30c7\u30eb\u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3059\u308b\u3053\u3068\u3067\u4eba\u6c17\u306e\u3042\u308b LLM \u306b\u30a2\u30af\u30bb\u30b9\u3067\u304d\u307e\u3059\u3002\u305d\u3057\u3066\u304a\u597d\u307f\u306e\u30e2\u30c7\u30eb\u3092\u4f7f\u3063\u3066\u30c6\u30ad\u30b9\u30c8\u89e3\u6790\u3084\u751f\u6210\u3092\u884c\u3046\u3053\u3068\u304c\u3067\u304d\u307e\u3059\u3002<br data-start=\"350\" data-end=\"353\" \/>LLM \u3068 MATLAB \u3092\u9023\u643a\u3055\u305b\u308b\u305f\u3081\u306e\u30b3\u30fc\u30c9\u306f\u3001<a href=\"https:\/\/jp.mathworks.com\/matlabcentral\/fileexchange\/163796-large-language-models-llms-with-matlab\">Large Language Models (LLMs) with MATLAB \u30ea\u30dd\u30b8\u30c8\u30ea<\/a> \u3067\u516c\u958b\u3055\u308c\u3066\u3044\u307e\u3059\u3002<\/p>\n<h6><\/h6>\n<p class=\"\" data-start=\"505\" data-end=\"529\">LLM \u306b\u30a2\u30af\u30bb\u30b9\u3059\u308b\u65b9\u6cd5\u306f\u4e3b\u306b\u4ee5\u4e0b\u306e3\u3064\u3067\u3059\uff1a<\/p>\n<ol data-start=\"531\" data-end=\"674\">\n<li class=\"\" data-start=\"531\" data-end=\"592\">\n<p class=\"\" data-start=\"534\" data-end=\"592\"><strong data-start=\"534\" data-end=\"566\">OpenAI\u00ae Chat Completions API<\/strong>\uff08ChatGPT \u306e\u30d9\u30fc\u30b9\u3068\u306a\u308b API\uff09\u3068\u63a5\u7d9a<\/p>\n<\/li>\n<li class=\"\" data-start=\"593\" data-end=\"627\">\n<p class=\"\" data-start=\"596\" data-end=\"627\"><strong data-start=\"596\" data-end=\"607\">Ollama\u2122<\/strong> \u3092\u5229\u7528\u3057\u305f\u30ed\u30fc\u30ab\u30eb LLM \u306e\u6d3b\u7528<\/p>\n<\/li>\n<li class=\"\" data-start=\"628\" data-end=\"674\">\n<p class=\"\" data-start=\"631\" data-end=\"674\"><strong data-start=\"631\" data-end=\"657\">Azure\u00ae OpenAI Services<\/strong> \u3092\u901a\u3058\u305f\u30af\u30e9\u30a6\u30c9\u30d9\u30fc\u30b9\u306e\u5229\u7528<\/p>\n<\/li>\n<\/ol>\n<p class=\"\" data-start=\"676\" data-end=\"716\">\u305d\u308c\u305e\u308c\u306e\u8a2d\u5b9a\u65b9\u6cd5\u3084\u4f7f\u7528\u4f8b\u306b\u3064\u3044\u3066\u306f\u3001\u4ee5\u4e0b\u306e\u30d6\u30ed\u30b0\u8a18\u4e8b\u3067\u8a73\u3057\u304f\u7d39\u4ecb\u3055\u308c\u3066\u3044\u307e\u3059\uff1a<\/p>\n<ul data-start=\"718\" data-end=\"929\" data-is-last-node=\"\" data-is-only-node=\"\">\n<li class=\"\" data-start=\"718\" data-end=\"824\">\n<p class=\"\" data-start=\"720\" data-end=\"824\"><a href=\"https:\/\/blogs.mathworks.com\/japan-community\/2024\/02\/06\/large-language-models-with-matlab-jp\/?from=jp\" target=\"_new\" rel=\"noopener\" data-start=\"720\" data-end=\"824\">MATLAB \u3067\u5927\u898f\u6a21\u8a00\u8a9e\u30e2\u30c7\u30eb<\/a><\/p>\n<\/li>\n<li data-start=\"718\" data-end=\"824\"><a href=\"https:\/\/blogs.mathworks.com\/deep-learning\/2024\/07\/09\/local-llms-with-matlab\/\" target=\"_new\" rel=\"noopener\" data-start=\"827\" data-end=\"929\" data-is-last-node=\"\">Local LLMs with MATLAB<\/a><\/li>\n<\/ul>\n<h6><\/h6>\n<p><img decoding=\"async\" loading=\"lazy\" class=\"alignnone wp-image-17285 \" src=\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2025\/04\/llms_with_matlab_repo.png\" alt=\"File Exchange repository: Large Language Models (LLMs) with MATLAB\" width=\"733\" height=\"514\" \/><\/p>\n<h6><\/h6>\n<p><em>File Exchange Repository: Large Language Models (LLMs) with MATLAB<\/em><\/p>\n<h6><\/h6>\n<p>&nbsp;<\/p>\n<h6><\/h6>\n<p style=\"font-size: 20px; color: #c04c0b;\"><strong>\u7570\u5e38\u5206\u985e\u306e\u30b3\u30fc\u30c9\u3092\u305f\u3081\u3057\u3066\u307f\u3088\u3046<\/strong><\/p>\n<p class=\"\" data-start=\"21\" data-end=\"110\">\u79c1\u306e\u540c\u50da\u3067\u3042\u308b<a href=\"https:\/\/www.linkedin.com\/in\/takuji-fukumoto-8a3013114\/\" class=\"external\" rel=\"nofollow\" target=\"_blank\">\u798f\u672c\u62d3\u53f8 (Takuji Fukumoto)<\/a> \u304c\u3001\u7570\u5e38\u5206\u985e\u3092\u5b9f\u884c\u3059\u308b\u305f\u3081\u306e\u30b3\u30fc\u30c9\u3092\u307e\u3068\u3081\u305f GitHub \u30ea\u30dd\u30b8\u30c8\u30ea: <a href=\"https:\/\/github.com\/matlab-deep-learning\/zero-shot-anomaly-classification-with-EfficientAD-and-LLM\" class=\"external\" rel=\"nofollow\" target=\"_blank\">Zero-shot anomaly classification with EfficientAD and LLM<\/a> \u304b\u3089\u30a2\u30af\u30bb\u30b9\u3067\u304d\u307e\u3059\u3002<\/p>\n<p class=\"\" data-start=\"246\" data-end=\"338\">\u3059\u3050\u306b\u8a66\u3057\u3066\u307f\u305f\u3044\u5834\u5408\u306f\u3001\u305d\u306e\u307e\u307e\u30ea\u30dd\u30b8\u30c8\u30ea\u3092\u78ba\u8a8d\u3057\u3066\u304f\u3060\u3055\u3044\u3002\u3053\u306e\u3042\u3068\u30ea\u30dd\u30b8\u30c8\u30ea\u306b\u542b\u307e\u308c\u308b\u4e3b\u8981\u306a\u30b9\u30c6\u30c3\u30d7\u3092\u9806\u3092\u8ffd\u3063\u3066\u89e3\u8aac\u3057\u3066\u3044\u304d\u307e\u3059\u3002<\/p>\n<h3 class=\"\" data-start=\"0\" data-end=\"15\">\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u306e\u8aad\u307f\u8fbc\u307f<\/h3>\n<p class=\"\" data-start=\"17\" data-end=\"114\">\u307e\u305a\u306f\u3001\u7570\u5e38\u5206\u985e\u3092\u884c\u3046\u753b\u50cf\u30c7\u30fc\u30bf\u3092\u7528\u610f\u3057\u3001\u305d\u308c\u3092 <strong data-start=\"42\" data-end=\"60\">ImageDatastore<\/strong> \u306b\u683c\u7d0d\u3057\u307e\u3059\u3002<code data-start=\"68\" data-end=\"84\">ImageDatastore<\/code> \u3092\u4f7f\u3046\u3053\u3068\u3067\u3001\u5927\u91cf\u306e\u753b\u50cf\u3092\u52b9\u7387\u3088\u304f\u6271\u3044\u306a\u304c\u3089\u51e6\u7406\u3067\u304d\u307e\u3059\u3002<\/p>\n<pre>dataDir = \"CAEdemo\";\r\ndownloadData(dataDir)\r\n\r\nImgsize = [720 1280]; % Image size\r\nBlockSize = [256 256]; % Input size of Network\r\nnumPatch =8;\r\n\r\nimds = imageDatastore(fullfile(dataDir,dataDir,\"trainingimage\"),IncludeSubfolders=true);\r\nimdsTrain = subset(imds,1:38);\r\nimdsTraint = transform(imdsTrain,@(x1) transformFcn(x1,BlockSize,numPatch));\r\n<\/pre>\n<h6><\/h6>\n<h3 class=\"\" data-start=\"0\" data-end=\"26\">\u5b66\u7fd2\u7528\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u304b\u3089\u30b5\u30f3\u30d7\u30eb\u753b\u50cf\u3092\u8868\u793a\u3059\u308b<\/h3>\n<p class=\"\" data-start=\"28\" data-end=\"111\">\u307e\u305a\u306f\u5b66\u7fd2\u7528\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u306b\u542b\u307e\u308c\u308b\u753b\u50cf\u306e\u4e00\u90e8\u3092\u8868\u793a\u3057\u3066\u307f\u307e\u3057\u3087\u3046\u3002\u3053\u308c\u3089\u306f\u300c\u5e03\u5730\uff08fabric\uff09\u300d\u306e\u753b\u50cf\u3067\u3001\u4e00\u898b\u3059\u308b\u3068\u7570\u5e38\u306e\u6709\u7121\u3092\u5224\u65ad\u3059\u308b\u306e\u306f\u975e\u5e38\u306b\u56f0\u96e3\u3067\u3059\u3002<\/p>\n<pre>numTrainImages = numel(imdsTrain.Files);\r\nidx = randperm(numTrainImages,16);\r\nI = imtile(imdsTrain,Frames=idx,GridSize=[4 4],BorderSize=10);\r\nfigure\r\nimshow(I)\r\n<\/pre>\n<h6><\/h6>\n<p><img decoding=\"async\" loading=\"lazy\" class=\"alignnone wp-image-17288 \" src=\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2025\/04\/input_images.png\" alt=\"Fabric images used as input for anomaly classification\" width=\"670\" height=\"385\" \/><\/p>\n<h6><\/h6>\n<p>&nbsp;<\/p>\n<h6><\/h6>\n<h3 class=\"\" data-start=\"0\" data-end=\"20\">\u7570\u5e38\u5206\u985e\u5668\u306e\u30d5\u30a1\u30a4\u30f3\u30c1\u30e5\u30fc\u30cb\u30f3\u30b0<\/h3>\n<p class=\"\" data-start=\"22\" data-end=\"176\">\u6b21\u306b\u3001<code data-start=\"25\" data-end=\"53\">efficientADAnomalyDetector<\/code> \u30aa\u30d6\u30b8\u30a7\u30af\u30c8\u3092\u4f7f\u7528\u3057\u3066\u3001EfficientAD \u7570\u5e38\u5206\u985e\u5668\u3092\u4f5c\u6210\u3057\u307e\u3059\u3002\u3053\u3053\u3067\u306f\u3001<code data-start=\"95\" data-end=\"116\">UseGlobalAnomalyMap<\/code> \u5f15\u6570\u3092 <code data-start=\"121\" data-end=\"128\">false<\/code> \u306b\u8a2d\u5b9a\u3057\u3066\u3001\u30aa\u30fc\u30c8\u30a8\u30f3\u30b3\u30fc\u30c0\u30fc\u3092\u7121\u52b9\u306b\u3057\u3001\u30b9\u30c1\u30e5\u30fc\u30c7\u30f3\u30c8\u30fb\u30c6\u30a3\u30fc\u30c1\u30e3\u30fc\u30e2\u30c7\u30eb\u306e\u307f\u3092\u4f7f\u7528\u3057\u307e\u3059\u3002<\/p>\n<pre>untrainedDetector = efficientADAnomalyDetector(Network=\"pdn-medium\",UseGlobalAnomalyMap=false);\r\n<\/pre>\n<h6><\/h6>\n<h3 class=\"\" data-start=\"0\" data-end=\"23\">\u8a13\u7df4\u30aa\u30d7\u30b7\u30e7\u30f3\u306e\u6307\u5b9a\u3068\u7570\u5e38\u691c\u51fa\u5668\u306e\u8a13\u7df4<\/h3>\n<p class=\"\" data-start=\"25\" data-end=\"110\">EfficientAD \u7570\u5e38\u691c\u51fa\u5668\u3092\u8a13\u7df4\u3059\u308b\u305f\u3081\u306b\u306f\u3001\u8a13\u7df4\u30aa\u30d7\u30b7\u30e7\u30f3\u3092\u6307\u5b9a\u3057\u3001\u305d\u306e\u5f8c\u306b\u30e2\u30c7\u30eb\u3092\u5b66\u7fd2\u3055\u305b\u307e\u3059\u3002\u4ee5\u4e0b\u306e\u30b3\u30fc\u30c9\u3067\u8a13\u7df4\u30aa\u30d7\u30b7\u30e7\u30f3\u3092\u8a2d\u5b9a\u3057\u3001\u7570\u5e38\u691c\u51fa\u5668\u3092\u8a13\u7df4\u3057\u307e\u3059\u3002<\/p>\n<pre>maximumEpochs=100;\r\noptions = trainingOptions(\"adam\", ...\r\n    InitialLearnRate=1e-4, ...\r\n    L2Regularization=1e-5, ...\r\n    LearnRateSchedule=\"piecewise\", ...\r\n    LearnRateDropPeriod=floor(0.95*maximumEpochs), ...\r\n    LearnRateDropFactor=0.1, ...\r\n    MaxEpochs=maximumEpochs, ...\r\n    VerboseFrequency=2, ...\r\n    MiniBatchSize=1, ...\r\n    Shuffle=\"every-epoch\", ...\r\n    ResetInputNormalization=true,...\r\n    Plots=\"training-progress\");\r\n\r\ndetector = trainEfficientADAnomalyDetector(imdsTraint,untrainedDetector,options,MapNormalizationDataRatio=0.2);\r\n<\/pre>\n<h6><\/h6>\n<p>\u3053\u308c\u3067\u3001\u7570\u5e38\u5206\u985e\u5668\u306f\u76ee\u7684\u306e\u30bf\u30b9\u30af\u306b\u4f7f\u7528\u3059\u308b\u6e96\u5099\u304c\u6574\u3044\u307e\u3057\u305f\u3002<\/p>\n<h6><\/h6>\n<p style=\"font-size: 18px;\"><strong>LLM \u306b\u3088\u308b\u5206\u985e\u306e\u5f37\u5316<\/strong><\/p>\n<p>\u3053\u3053\u304b\u3089 GPT \u30e2\u30c7\u30eb\u3092\u6d3b\u7528\u3057\u3066\u7570\u5e38\u5206\u985e\u306e\u7d50\u679c\u3092\u5f37\u5316\u3057\u307e\u3059\u3002LLM \u306b\u3088\u308a\u3001\u7570\u5e38\u5206\u985e\u306e\u8aac\u660e\u3092\u81ea\u7136\u8a00\u8a9e\u3067\u751f\u6210\u3057\u305f\u308a\u3001\u7570\u5e38\u306e\u539f\u56e0\u3084\u63a8\u5968\u5bfe\u5fdc\u7b56\u3092\u63d0\u6848\u3057\u305f\u308a\u3067\u304d\u307e\u3059\u3002MATLAB \u304b\u3089 OpenAI \u306e GPT \u30e2\u30c7\u30eb\u3092\u4f7f\u7528\u3059\u308b\u306b\u306f API key \u304c\u5fc5\u8981\u3067\u3059\u304c\u3001\u8a73\u7d30\u306f <a href=\"https:\/\/openai.com\/blog\/openai-api\" class=\"external\" rel=\"nofollow\" target=\"_blank\">ChatGPT API<\/a>\u00a0\u3092\u78ba\u8a8d\u304f\u3060\u3055\u3044\u3002<\/p>\n<p>\u307e\u305a OpenAI Chat \u30aa\u30d6\u30b8\u30a7\u30af\u30c8\u3092\u521d\u671f\u5316\u3057\u307e\u3059\u3002<\/p>\n<pre>chat = openAIChat(\"You are an AI assistant.\",ApiKey=my_key);\r\n<\/pre>\n<h6><\/h6>\n<p>&nbsp;<\/p>\n<p class=\"\" data-start=\"28\" data-end=\"127\">EfficientAD \u7570\u5e38\u691c\u51fa\u5668\u306b\u3088\u3063\u3066\u751f\u6210\u3055\u308c\u305f\u7570\u5e38\u30de\u30c3\u30d7\u3068\u3001LLM \u306b\u3088\u3063\u3066\u751f\u6210\u3055\u308c\u305f\u81ea\u7136\u8a00\u8a9e\u306e\u8aac\u660e\u3092\u7d44\u307f\u5408\u308f\u305b\u3066\u3001\u753b\u50cf\u3092\u8868\u793a\u3057\u306a\u304c\u3089\u7570\u5e38\u5206\u985e\u7d50\u679c\u3092\u89e3\u8aac\u3057\u307e\u3059\u3002\u4ee5\u4e0b\u306e\u6d41\u308c\u3067\u3053\u308c\u3092\u5b9f\u884c\u3067\u304d\u307e\u3059\u3002<\/p>\n<ol data-start=\"129\" data-end=\"299\">\n<li class=\"\" data-start=\"129\" data-end=\"228\">\n<p class=\"\" data-start=\"132\" data-end=\"228\"><strong data-start=\"132\" data-end=\"144\">\u7570\u5e38\u30de\u30c3\u30d7\u306e\u751f\u6210<\/strong>\uff1aEfficientAD \u3092\u4f7f\u3063\u3066\u7570\u5e38\u30de\u30c3\u30d7\u3092\u751f\u6210\u3057\u307e\u3059\u3002\u7570\u5e38\u30de\u30c3\u30d7\u306f\u3001\u753b\u50cf\u306e\u5404\u30d4\u30af\u30bb\u30eb\u306e\u7570\u5e38\u5ea6\u5408\u3044\u3092\u793a\u3059\u3082\u306e\u3067\u3001\u3069\u306e\u90e8\u5206\u304c\u901a\u5e38\u306e\u30d1\u30bf\u30fc\u30f3\u3068\u7570\u306a\u308b\u304b\u3092\u8996\u899a\u7684\u306b\u793a\u3057\u307e\u3059\u3002<\/p>\n<\/li>\n<li class=\"\" data-start=\"229\" data-end=\"299\">\n<p class=\"\" data-start=\"232\" data-end=\"299\"><strong data-start=\"232\" data-end=\"254\">LLM \u3092\u4f7f\u7528\u3057\u305f\u81ea\u7136\u8a00\u8a9e\u306b\u3088\u308b\u89e3\u8aac<\/strong>\uff1a\u751f\u6210\u3055\u308c\u305f\u7570\u5e38\u30de\u30c3\u30d7\u306b\u57fa\u3065\u3044\u3066\u3001LLM \u304c\u7570\u5e38\u306e\u539f\u56e0\u3084\u5206\u985e\u7d50\u679c\u306b\u3064\u3044\u3066\u8aac\u660e\u3092\u751f\u6210\u3057\u307e\u3059\u3002<\/p>\n<\/li>\n<\/ol>\n<pre>alltext = [];\r\nimdsllm = imageDatastore(\"cropped_image\",IncludeSubfolders=true);\r\n\r\nfor k =1:numel(imdsllm.Files)\r\n    % Create context and query prompts\r\n    messages = messageHistory;\r\n    context = \"You are inspecting the blue polka dot fabric as shown in this image file. This file is a normal data.\" + ...\r\n        \"It has a tested image on its left and a mask image showing where there is abnormal part on its right\" + ...\r\n        \"If the test image is normal, the right side is completely black and no white area.\";\r\n    messages = addUserMessageWithImages(messages,context,string(imdsllm.Files{6}));\r\n\r\n    image_path = imdsllm.Files{k};\r\n    Query = \"This image is the one that was determined to be abnormal.\" + ...\r\n        \"It has a tested image on its left and a mask image showing where there is abnormal part on its right\" + ...\r\n        \"Referencing white area of the mask image describe What anomalies the image on your left contains with in 10 words\" + ...\r\n        \"It is possible that a foreign substance is mixed in, something different in color exist or there is some kind of stain on the fabric.\"+ ...\r\n        \"If the corresponding mask on your right is black and not white, There are no stains and foreign substances. Answer 'This is normal'\";\r\n    messages = addUserMessageWithImages(messages,Query,string(image_path));\r\n    \r\n    % Generate a response\r\n    [txt,response] = generate(chat,messages,MaxNumTokens=4096,TimeOut=500);\r\n    \r\n    % Plot anomaly classification results enhanced by LLM\r\n    figure(f), subplot(6,1,k), imshow(imoutall{k});\r\n    title(txt,'FontSize',10);shg\r\nend\r\n<\/pre>\n<h6><\/h6>\n<p><img decoding=\"async\" loading=\"lazy\" class=\"alignnone wp-image-17291 \" src=\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2025\/04\/classification_result.png\" alt=\"LLM-enhanced anomaly classification results\" width=\"514\" height=\"627\" \/><\/p>\n<h6><\/h6>\n<p>&nbsp;<\/p>\n<h6><\/h6>\n<p style=\"font-size: 20px; color: #c04c0b;\">\u4e3b\u306a\u30dd\u30a4\u30f3\u30c8<\/p>\n<ol>\n<li>EfficientAD \u30e2\u30c7\u30eb \u306f\u3001\u8efd\u91cf\u306a\u30a2\u30fc\u30ad\u30c6\u30af\u30c1\u30e3\u3068\u5f37\u529b\u306a\u610f\u5473\u7684\u7406\u89e3\u3092\u7d44\u307f\u5408\u308f\u305b\u308b\u3053\u3068\u3067\u3001\u8fc5\u901f\u3067\u9ad8\u7cbe\u5ea6\u304b\u3064\u9ad8\u901f\u306a\u7570\u5e38\u5206\u985e\u3092\u5b9f\u73fe\u3057\u307e\u3059\u3002\u3053\u308c\u306b\u3088\u308a\u3001\u30ea\u30a2\u30eb\u30bf\u30a4\u30e0\u306e\u7523\u696d\u30a2\u30d7\u30ea\u30b1\u30fc\u30b7\u30e7\u30f3\u306b\u6700\u9069\u3067\u3059\u3002<\/li>\n<li>LLM\uff08\u5927\u898f\u6a21\u8a00\u8a9e\u30e2\u30c7\u30eb\uff09 \u306f\u3001\u7570\u5e38\u5206\u985e\u7d50\u679c\u3092\u81ea\u7136\u8a00\u8a9e\u3067\u8aac\u660e\u306b\u5909\u63db\u3057\u3001\u904e\u53bb\u306e\u30c7\u30fc\u30bf\u306b\u57fa\u3065\u3044\u3066\u6f5c\u5728\u7684\u306a\u539f\u56e0\u3092\u63d0\u6848\u3057\u307e\u3059\u3002\u307e\u305f\u3001\u7570\u5e38\u30ec\u30dd\u30fc\u30c8\u306e\u81ea\u52d5\u751f\u6210\u3092\u652f\u63f4\u3057\u3001\u5206\u985e\u7d50\u679c\u3092\u3088\u308a\u89e3\u91c8\u53ef\u80fd\u3067\u5b9f\u884c\u53ef\u80fd\u306a\u3082\u306e\u306b\u3057\u307e\u3059\u3002<\/li>\n<li>MATLAB \u306f\u3001AI \u6280\u8853\u3084\u30c7\u30fc\u30bf\u524d\u51e6\u7406\u6280\u8853\u3092\u7d71\u5408\u3067\u304d\u308b\u74b0\u5883\u3092\u63d0\u4f9b\u3057\u3001\u30a8\u30f3\u30c9\u30c4\u30fc\u30a8\u30f3\u30c9\u306e\u7570\u5e38\u5206\u985e\u30ef\u30fc\u30af\u30d5\u30ed\u30fc\u3092\u69cb\u7bc9\u3067\u304d\u307e\u3059\u3002\u3053\u306e\u7d71\u5408\u7684\u306a\u30a2\u30d7\u30ed\u30fc\u30c1\u306b\u3088\u308a\u3001\u958b\u767a\u3001\u30c7\u30d7\u30ed\u30a4\u3001\u8aac\u660e\u6027\u306e\u5411\u4e0a\u304c\u7c21\u7d20\u5316\u3055\u308c\u307e\u3059\u3002<\/li>\n<\/ol>\n<h6><\/h6>\n","protected":false},"excerpt":{"rendered":"<div class=\"overview-image\"><img decoding=\"async\"  class=\"img-responsive\" src=\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2025\/04\/anomaly_classification_intro.png\" onError=\"this.style.display ='none';\" \/><\/div>\n<p>\u203b\u3053\u306e\u6295\u7a3f\u306f 2025 \u5e74 4 \u6708 21 \u65e5\u306b Artificial Intelligence \u3078 \u6295\u7a3f\u3055\u308c\u305f\u3082\u306e\u306e\u6284\u8a33\u3067\u3059\u3002<\/p>\n<p>\u753b\u50cf\u5185\u306e\u4e88\u671f\u3057\u306a\u3044\u30d1\u30bf\u30fc\u30f3\u306b\u3001\u610f\u5473\u306e\u3042\u308b\u30ab\u30c6\u30b4\u30ea\u3084\u8aac\u660e\u3092\u4ed8\u3051\u308b\u3053\u3068\uff08= \u753b\u50cf\u306e\u7570\u5e38\u5206\u985e\uff09\u306f\u3001\u54c1\u8cea\u7ba1\u7406\u306e\u81ea\u52d5\u5316\u3084\u533b\u7642\u8a3a\u65ad\u306e\u9ad8\u5ea6\u5316\u306b\u5f79\u7acb\u3061\u307e\u3059\u3002\u672c\u30d6\u30ed\u30b0\u3067\u306f\u3001\u753b\u50cf\u30c7\u30fc\u30bf\u5411\u3051\u306e\u7570\u5e38\u5206\u985e\u30e2\u30c7\u30eb\u3067\u3042\u308b EfficientAD&#8230; <a class=\"read-more\" href=\"https:\/\/blogs.mathworks.com\/japan-community\/2025\/05\/01\/llm-enhanced-anomaly-classification-for-images-jp\/\">read more >><\/a><\/p>\n","protected":false},"author":159,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":[],"categories":[182,61],"tags":[],"_links":{"self":[{"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/posts\/13325"}],"collection":[{"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/users\/159"}],"replies":[{"embeddable":true,"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/comments?post=13325"}],"version-history":[{"count":15,"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/posts\/13325\/revisions"}],"predecessor-version":[{"id":13445,"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/posts\/13325\/revisions\/13445"}],"wp:attachment":[{"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/media?parent=13325"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/categories?post=13325"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/tags?post=13325"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}