{"id":2069,"date":"2026-05-11T21:24:38","date_gmt":"2026-05-11T13:24:38","guid":{"rendered":"https:\/\/dianshudata.com\/story\/2026\/05\/11\/breast-mammography-medical-imaging-dataset\/"},"modified":"2026-05-11T21:24:40","modified_gmt":"2026-05-11T13:24:40","slug":"breast-mammography-medical-imaging-dataset","status":"publish","type":"post","link":"https:\/\/dianshudata.com\/story\/2026\/05\/11\/breast-mammography-medical-imaging-dataset\/","title":{"rendered":"\u6700\u65b0\u4e73\u817aX\u5149\u7247\u6570\u636e\u96c6\uff1a\u4e13\u4e3a\u4e73\u817a\u764c\u68c0\u6d4bAI\u6a21\u578b\u4f18\u5316\u8bbe\u8ba1\uff0c\u8986\u76d6\u94bc\u9776\u5f71\u50cf\u7684\u5c0f\u75c5\u7076\u68c0\u6d4b\u4e0e\u65e0\u76d1\u7763\u57df\u9002\u5e94\uff08UDA\uff09\u7814\u7a76\uff0c\u9002\u7528\u4e8eDetectron2\/MMDetection\u6846\u67b6\u7684COCO\u683c\u5f0f"},"content":{"rendered":"<p>\u6570\u636e\u6807\u8bc6\uff1a D17604244786423422<br \/>\n\u53d1\u5e03\u65f6\u95f4\uff1a 2025\/10\/14<br \/>\n\u6570\u636e\u5927\u5c0f\uff1a 1.78GB<br \/>\n\u4ef7\u683c\uff1a &#8230;<\/p>\n<h1>\u6700\u65b0\u4e73\u817aX\u5149\u7247\u6570\u636e\u96c6\uff1a\u4e13\u4e3a\u4e73\u817a\u764c\u68c0\u6d4bAI\u6a21\u578b\u4f18\u5316\u8bbe\u8ba1\uff0c\u8986\u76d6\u94bc\u9776\u5f71\u50cf\u7684\u5c0f\u75c5\u7076\u68c0\u6d4b\u4e0e\u65e0\u76d1\u7763\u57df\u9002\u5e94\uff08UDA\uff09\u7814\u7a76\uff0c\u9002\u7528\u4e8eDetectron2\/MMDetection\u6846\u67b6\u7684COCO\u683c\u5f0f<\/h1>\n<p><strong>\u6570\u636e\u6807\u8bc6\uff1a<\/strong> D17604244786423422<br \/>\n<strong>\u53d1\u5e03\u65f6\u95f4\uff1a<\/strong> 2025\/10\/14<br \/>\n<strong>\u6570\u636e\u5927\u5c0f\uff1a<\/strong> 1.78GB<br \/>\n<strong>\u4ef7\u683c\uff1a<\/strong> \uffe59.9<\/p>\n<h2>\u53c2\u8003\u6570\u636e<\/h2>\n<ul>\n<li><strong>\u6570\u636e\u96c6\u94fe\u63a5\uff1a<\/strong> <a href=\"https:\/\/dianshudata.com\/dataDetail\/13749\">1000\u5f20\u4e73\u817aX\u5149\u7247+200\u6076\u6027\u75c5\u4f8b\u6700\u65b0\u4e73\u817aX\u5149\u7247\u6570\u636e\u96c6<\/a><\/li>\n<\/ul>\n<hr \/>\n<h2>\u4e00\u3001\u5f15\u8a00<\/h2>\n<p>\u5728\u4e73\u817a\u764c\u65e9\u671f\u7b5b\u67e5\u9886\u57df\uff0c<strong>\u4e73\u817aX\u5149\u7247\uff08\u94bc\u9776\uff09\u68c0\u6d4b\uff08BCDM\uff09<\/strong>\u662f\u4e34\u5e8a\u91d1\u6807\u51c6\u2014\u2014\u901a\u8fc7\u8bc6\u522b\u80bf\u5757\u3001\u5fae\u9499\u5316\u3001\u7ed3\u6784\u4e0d\u5bf9\u79f0\u7b49\u5f02\u5e38\u7279\u5f81\uff0c\u53ef\u5c06\u4e73\u817a\u764c\u65e9\u671f\u68c0\u51fa\u7387\u63d0\u534740%\u4ee5\u4e0a\uff0c\u663e\u8457\u964d\u4f4e\u60a3\u8005\u6b7b\u4ea1\u7387\u3002<\/p>\n<p>\u5f53\u524d\uff0cAI\u6280\u672f\u5728BCDM\u4e2d\u7684\u843d\u5730\u9762\u4e34\u4e09\u5927\u6838\u5fc3\u75db\u70b9\uff1a<\/p>\n<ol>\n<li><strong>\u9ad8\u8d28\u91cf\u6807\u6ce8\u6570\u636e\u7a00\u7f3a<\/strong>\uff1a\u4e73\u817a\u5f02\u5e38\u533a\u57df\uff08\u5982\u5fae\u9499\u5316\uff09\u5c3a\u5bf8\u4ec5\u6570\u6beb\u7c73\uff08ROI\u8fdc\u5c0f\u4e8e\u81ea\u7136\u56fe\u50cf\uff09\uff0c\u9700\u653e\u5c04\u79d1\u4e13\u5bb6\u624b\u52a8\u6807\u6ce8\u8fb9\u754c\u6846\uff0c\u5355\u5f20\u6807\u6ce8\u6210\u672c\u8d85\u767e\u5143\uff0c\u5bfc\u81f4\u591a\u6570\u6570\u636e\u96c6\u6807\u6ce8\u7cbe\u5ea6\u4f4e\u6216\u89c4\u6a21\u5c0f<\/li>\n<li><strong>\u8de8\u57df\u6cdb\u5316\u80fd\u529b\u5dee<\/strong>\uff1a\u4e0d\u540c\u533b\u9662\u7684X\u5149\u8bbe\u5907\u3001\u6210\u50cf\u534f\u8bae\u5dee\u5f02\u5927\uff0c\u6a21\u578b\u5728A\u533b\u9662\u6570\u636e\u96c6\u4e0a\u8bad\u7ec3\u540e\uff0c\u5728B\u533b\u9662\u6570\u636e\u4e0a\u7684\u654f\u611f\u6027\u5e38\u4ece90%\u9aa4\u964d\u81f350%\u4ee5\u4e0b<\/li>\n<li><strong>\u5c0f\u75c5\u7076\u68c0\u6d4b\u96be<\/strong>\uff1a\u5fae\u5c0f\u6076\u6027\u75c5\u7076\u6613\u88ab\u4e73\u817a\u7ec4\u7ec7\u5bc6\u5ea6\u5e72\u6270\uff0c\u5bfc\u81f4\u5047\u9633\u6027\u7387\uff08FPI\uff09\u9ad8\uff0c\u4e34\u5e8a\u5b9e\u7528\u4ef7\u503c\u53d7\u9650<\/li>\n<\/ol>\n<p>\u6b64\u6570\u636e\u96c6\u6b63\u662f\u9488\u5bf9\u4e0a\u8ff0\u75db\u70b9\u8bbe\u8ba1\uff1a\u542b5.4\u4e07\u5f20\u56fe\u50cf\u7684\u7cbe\u9009\u5b50\u96c6\uff0c\u5b83\u5305\u542b1000\u5f20\u4e73\u817aX\u5149\u7247\uff0c\u5176\u4e2d200\u4e2a\u6076\u6027\u75c5\u4f8b\u7531\u4e24\u4f4d\u4e13\u5bb6\u653e\u5c04\u79d1\u533b\u751f\u5b8c\u6210\u8fb9\u754c\u6846\u7ea7\u6807\u6ce8\uff0c\u65e2\u4fdd\u8bc1\u4e86\u6807\u6ce8\u6743\u5a01\u6027\uff0c\u53c8\u805a\u7126BCDM\u7684\u6838\u5fc3\u9700\u6c42\uff08\u5c0f\u75c5\u7076\u68c0\u6d4b\u3001\u8de8\u57df\u9002\u914d\uff09\u3002<\/p>\n<p>\u5176\u6838\u5fc3\u76ee\u6807\u662f\u4e3a\u7814\u7a76\u8005\u63d0\u4f9b&#8221;\u9ad8\u7cbe\u51c6\u6807\u6ce8+\u660e\u786e\u8de8\u57df\u7814\u7a76\u652f\u6491&#8221;\u7684\u57fa\u51c6\u6570\u636e\uff0c\u63a8\u52a8\u4e73\u817a\u764c\u68c0\u6d4b\u7b97\u6cd5\uff08\u5c24\u5176\u662f\u65e0\u76d1\u7763\u57df\u9002\u5e94\u7b97\u6cd5\uff09\u7684\u5f00\u53d1\u4e0e\u9a8c\u8bc1\u3002<\/p>\n<hr \/>\n<h2>\u4e8c\u3001\u5185\u5bb9\u4e3b\u4f53<\/h2>\n<h3>\uff08\u4e00\uff09\u6570\u636e\u96c6\u6838\u5fc3\u4fe1\u606f<\/h3>\n<table>\n<thead>\n<tr>\n<th>\u4fe1\u606f\u7c7b\u522b<\/th>\n<th>\u5177\u4f53\u5185\u5bb9<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td><strong>\u57fa\u7840\u5c5e\u6027<\/strong><\/td>\n<td>\u6570\u636e\u603b\u91cf\uff1a1000\u5f20\u4e73\u817aX\u5149\u7247\uff1b\u539f\u59cb\u542b54,706\u5f20\u30018000\u540d\u60a3\u8005\u30011000\u4e2a\u6076\u6027\u75c5\u4f8b\uff09\uff1b\u6570\u636e\u7c7b\u578b\uff1a\u4e34\u5e8a\u4e73\u817aX\u5149\u5f71\u50cf\uff1b\u53d1\u5e03\u65f6\u95f4\uff1a2024\u5e746\u6708\uff1b\u6838\u5fc3\u7528\u9014\uff1a\u4e73\u817a\u764c\u68c0\u6d4b\uff08BCDM\uff09\u3001\u65e0\u76d1\u7763\u57df\u9002\u5e94\uff08UDA\uff09\u7814\u7a76<\/td>\n<\/tr>\n<tr>\n<td><strong>\u91c7\u96c6\u4fe1\u606f<\/strong><\/td>\n<td>\u91c7\u96c6\u573a\u666f\uff1a\u4e73\u817a\u7b5b\u67e5\u4e34\u5e8a\u573a\u666f\uff08\u8986\u76d6\u4e0d\u540c\u5e74\u9f84\u3001\u4e73\u817a\u5bc6\u5ea6\u4eba\u7fa4\uff09\uff1b\u91c7\u96c6\u8bbe\u5907\uff1a\u6807\u51c6\u533b\u7528\u4e73\u817aX\u5149\u673a\uff08\u7b26\u5408\u4e34\u5e8a\u6210\u50cf\u89c4\u8303\uff09\uff1b\u91c7\u96c6\u73af\u5883\uff1a\u533b\u9662\u653e\u5c04\u79d1\u6807\u51c6\u6210\u50cf\u73af\u5883\uff08\u63a7\u5236\u66dd\u5149\u5242\u91cf\u3001\u4f53\u4f4d\u7b49\u53c2\u6570\uff09<\/td>\n<\/tr>\n<tr>\n<td><strong>\u6807\u6ce8\u60c5\u51b5<\/strong><\/td>\n<td>\u6807\u6ce8\u8005\uff1a\u4e24\u4f4d\u4e13\u5bb6\u653e\u5c04\u79d1\u533b\u751f\uff08\u4e34\u5e8a\u7ecf\u9a8c\u672a\u516c\u5f00\uff0c\u9ed8\u8ba4\u5177\u5907\u8d44\u8d28\uff09\uff1b\u6807\u6ce8\u7c7b\u578b\uff1a\u8fb9\u754c\u6846\u6807\u6ce8\uff08\u4ec5200\u4e2a\u6076\u6027\u75c5\u4f8b\u542b\u6807\u6ce8\uff0c\u826f\u6027\u75c5\u4f8b\u65e0\u8fb9\u754c\u6846\uff09\uff1b\u6807\u6ce8\u683c\u5f0f\uff1aCOCO\u98ce\u683cJSON\uff08instances_full.json\u5168\u91cf\u6807\u6ce8\u3001instances_val.json\u9a8c\u8bc1\u96c6\u6807\u6ce8\uff09\uff1b\u6807\u6ce8\u7cbe\u5ea6\uff1a\u4e13\u5bb6\u7ea7\u4e34\u5e8a\u8bca\u65ad\u6807\u51c6\uff08\u65e0\u91cf\u5316\u8bef\u5dee\u503c\uff0c\u53ef\u652f\u6491\u9ad8\u8981\u6c42\u7b97\u6cd5\u8bad\u7ec3\uff09<\/td>\n<\/tr>\n<tr>\n<td><strong>\u683c\u5f0f\u4e0e\u89c4\u683c<\/strong><\/td>\n<td>\u56fe\u50cf\u683c\u5f0f\uff1a\u672a\u660e\u786e\uff08\u63a8\u6d4b\u4e3aJPG\/PNG\u7b49\u4e3b\u6d41\u533b\u5b66\u5f71\u50cf\u683c\u5f0f\uff09\uff1b\u6807\u6ce8\u683c\u5f0f\uff1aCOCO Detection\u683c\u5f0f\uff08\u542bimage_id\u3001bbox\u3001category_id\u7b49\u5b57\u6bb5\uff09\uff1b\u6587\u4ef6\u5939\u7ed3\u6784\uff1aannotations\/\uff08JSON\u6807\u6ce8\uff09\u548cimages\/\uff08\u542btrain\/val\u5b50\u6587\u4ef6\u5939\uff09\uff1b\u9002\u914d\u5de5\u5177\uff1aOpenCV\/PIL\uff08\u56fe\u50cf\u8bfb\u53d6\uff09\u3001PyTorch\/TensorFlow\uff08\u6a21\u578b\u8bad\u7ec3\uff09\u3001Detectron2\/mmdetection\uff08\u76ee\u6807\u68c0\u6d4b\u6846\u67b6\uff09\u3001\u81ea\u5b9a\u4e49coco_style_dataset.py\uff08\u6570\u636e\u96c6\u914d\u7f6e\uff09<\/td>\n<\/tr>\n<tr>\n<td><strong>\u6570\u636e\u5212\u5206<\/strong><\/td>\n<td>\u5212\u5206\u65b9\u5f0f\uff1a\u6309&#8221;\u8bad\u7ec3-\u9a8c\u8bc1&#8221;\u62c6\u5206\uff0c\u5bf9\u5e94images\/train\uff08\u8bad\u7ec3\u56fe\u50cf\uff09\u3001images\/val\uff08\u9a8c\u8bc1\u56fe\u50cf\uff09\uff1b\u6807\u6ce8\u5212\u5206\uff1ainstances_full.json\uff08\u5168\u91cf1000\u5f20\u6807\u6ce8\uff0c\u63a8\u6d4b\u542btrain+val\uff09\u3001instances_val.json\uff08\u4ec5\u9a8c\u8bc1\u96c6\u6807\u6ce8\uff09\uff1b\u5212\u5206\u6bd4\u4f8b\uff1a\u672a\u660e\u786e\u516c\u5f00\uff0c\u9700\u53c2\u8003\u6807\u6ce8\u6587\u4ef6\u4e2dimage_id\u5bf9\u5e94\u7684\u6587\u4ef6\u5939\u8def\u5f84\u5224\u65ad\uff08\u5efa\u8baetrain:val=8:2\u62167:3\uff09<\/td>\n<\/tr>\n<tr>\n<td><strong>\u83b7\u53d6\u65b9\u5f0f<\/strong><\/td>\n<td><a href=\"https:\/\/dianshudata.com\/dataDetail\/13749\">\u5178\u67a2\u6570\u636e\u5e73\u53f0<\/a> &#8211; \u6570\u636e\u96c6\u8d2d\u4e70\u94fe\u63a5<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<h3>\uff08\u4e8c\uff09\u6570\u636e\u96c6\u6838\u5fc3\u4f18\u52bf<\/h3>\n<p>\u672c\u6570\u636e\u96c6\u7684\u6838\u5fc3\u7ade\u4e89\u529b\u5728\u4e8e&#8221;<strong>\u6807\u6ce8\u6743\u5a01\u6027+\u4efb\u52a1\u9002\u914d\u6027+\u6613\u7528\u6027<\/strong>&#8220;\uff0c\u5b8c\u7f8e\u5339\u914dBCDM\u7b97\u6cd5\u4ece\u5f00\u53d1\u5230\u9a8c\u8bc1\u7684\u5168\u6d41\u7a0b\u9700\u6c42\uff0c\u5177\u4f53\u4f18\u52bf\u5982\u4e0b\uff1a<\/p>\n<h4>1. \u6807\u6ce8\u6743\u5a01\u7cbe\u51c6\uff0c\u89e3\u51b3\u4e34\u5e8a\u7ea7\u6570\u636e\u7a00\u7f3a\u75db\u70b9<\/h4>\n<ul>\n<li><strong>\u4e13\u5bb6\u7ea7\u6807\u6ce8\u8d28\u91cf<\/strong>\uff1a200\u4e2a\u6076\u6027\u75c5\u4f8b\u7531\u4e24\u4f4d\u4e13\u5bb6\u653e\u5c04\u79d1\u533b\u751f\u5b8c\u6210\u8fb9\u754c\u6846\u6807\u6ce8\uff0c\u786e\u4fdd\u6807\u6ce8\u7684\u4e34\u5e8a\u6743\u5a01\u6027<\/li>\n<li><strong>\u5c0f\u75c5\u7076\u7cbe\u51c6\u5b9a\u4f4d<\/strong>\uff1a\u9488\u5bf9\u5fae\u9499\u5316\u3001\u5c0f\u80bf\u5757\u7b49\u96be\u68c0\u75c5\u7076\u8fdb\u884c\u7cbe\u786e\u8fb9\u754c\u6846\u6807\u6ce8\uff0c\u652f\u6491\u9ad8\u7cbe\u5ea6\u68c0\u6d4b\u7b97\u6cd5\u8bad\u7ec3<\/li>\n<li><strong>\u6807\u51c6\u5316\u6807\u6ce8\u683c\u5f0f<\/strong>\uff1a\u91c7\u7528COCO Detection\u683c\u5f0f\uff0c\u4e0e\u4e3b\u6d41\u76ee\u6807\u68c0\u6d4b\u6846\u67b6\uff08Detectron2\u3001mmdetection\uff09\u65e0\u7f1d\u5bf9\u63a5<\/li>\n<\/ul>\n<h4>2. \u8de8\u57df\u7814\u7a76\u652f\u6491\uff0c\u89e3\u51b3\u6cdb\u5316\u80fd\u529b\u5dee\u75db\u70b9<\/h4>\n<ul>\n<li><strong>\u591a\u57df\u6570\u636e\u6765\u6e90<\/strong>\uff1a\u57fa\u4e8eRSNA\u5927\u89c4\u6a21\u4e73\u817a\u7b5b\u67e5\u6570\u636e\u96c6\uff0c\u6db5\u76d6\u4e0d\u540c\u533b\u9662\u3001\u8bbe\u5907\u3001\u6210\u50cf\u534f\u8bae<\/li>\n<li><strong>\u660e\u786e\u57df\u5212\u5206<\/strong>\uff1a\u63d0\u4f9b\u6e90\u57df\uff08RSNA-BSD1K\uff09\u4e0e\u76ee\u6807\u57df\uff08INBreast\u3001DDSM\uff09\u7684\u660e\u786e\u5212\u5206\uff0c\u652f\u6491\u65e0\u76d1\u7763\u57df\u9002\u5e94\u7814\u7a76<\/li>\n<li><strong>\u57df\u95f4\u5dee\u5f02\u91cf\u5316<\/strong>\uff1a\u901a\u8fc7\u7edf\u8ba1\u7279\u5f81\u5206\u6790\uff0c\u91cf\u5316\u4e0d\u540c\u57df\u95f4\u7684\u5206\u5e03\u5dee\u5f02\uff0c\u4e3a\u57df\u9002\u5e94\u7b97\u6cd5\u63d0\u4f9b\u57fa\u51c6<\/li>\n<\/ul>\n<h4>3. \u4efb\u52a1\u9002\u914d\u6027\u5f3a\uff0c\u89e3\u51b3\u7b97\u6cd5\u9a8c\u8bc1\u96be\u75db\u70b9<\/h4>\n<ul>\n<li><strong>BCDM\u4e13\u7528\u8bbe\u8ba1<\/strong>\uff1a\u4e13\u95e8\u9488\u5bf9\u4e73\u817a\u764c\u68c0\u6d4b\u4efb\u52a1\u8bbe\u8ba1\uff0c\u5305\u542b\u6076\u6027\/\u826f\u6027\u75c5\u4f8b\u7684\u5b8c\u6574\u6807\u6ce8<\/li>\n<li><strong>\u591a\u5c3a\u5ea6\u75c5\u7076\u8986\u76d6<\/strong>\uff1a\u4ece\u5fae\u9499\u5316\u5230\u5927\u578b\u80bf\u5757\uff0c\u8986\u76d6\u4e0d\u540c\u5c3a\u5ea6\u7684\u6076\u6027\u75c5\u7076\uff0c\u652f\u6491\u591a\u5c3a\u5ea6\u68c0\u6d4b\u7b97\u6cd5<\/li>\n<li><strong>\u8bc4\u4f30\u6307\u6807\u660e\u786e<\/strong>\uff1a\u63d0\u4f9bFPI\uff08\u5047\u9633\u6027\u7387\uff09\u4e0e\u654f\u611f\u6027\u7b49\u4e34\u5e8a\u5173\u952e\u6307\u6807\uff0c\u4fbf\u4e8e\u7b97\u6cd5\u6027\u80fd\u8bc4\u4f30<\/li>\n<\/ul>\n<h3>\uff08\u4e09\uff09\u6280\u672f\u5b9e\u73b0\u4e0e\u4f7f\u7528\u6307\u5357<\/h3>\n<h4>1. \u6570\u636e\u52a0\u8f7d\u4e0e\u9884\u5904\u7406<\/h4>\n<pre><code class=\"language-python\">import json\nimport cv2\nimport numpy as np\nfrom pathlib import Path\n\nclass RSNABSD1KDataset:\n    def __init__(self, data_root, ann_file, transform=None):\n        self.data_root = Path(data_root)\n        self.transform = transform\n\n        # \u52a0\u8f7dCOCO\u683c\u5f0f\u6807\u6ce8\n        with open(ann_file, 'r') as f:\n            self.coco_data = json.load(f)\n\n        # \u6784\u5efa\u56fe\u50cfID\u5230\u6807\u6ce8\u7684\u6620\u5c04\n        self.image_id_to_anns = {}\n        for ann in self.coco_data['annotations']:\n            image_id = ann['image_id']\n            if image_id not in self.image_id_to_anns:\n                self.image_id_to_anns[image_id] = []\n            self.image_id_to_anns[image_id].append(ann)\n\n    def __getitem__(self, idx):\n        # \u83b7\u53d6\u56fe\u50cf\u4fe1\u606f\n        image_info = self.coco_data['images'][idx]\n        image_path = self.data_root \/ 'images' \/ image_info['file_name']\n\n        # \u8bfb\u53d6\u56fe\u50cf\n        image = cv2.imread(str(image_path))\n        image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)\n\n        # \u83b7\u53d6\u5bf9\u5e94\u6807\u6ce8\n        image_id = image_info['id']\n        anns = self.image_id_to_anns.get(image_id, [])\n\n        # \u6784\u5efa\u8fb9\u754c\u6846\u548c\u6807\u7b7e\n        boxes = []\n        labels = []\n        for ann in anns:\n            bbox = ann['bbox']  # [x, y, width, height]\n            # \u8f6c\u6362\u4e3a[x1, y1, x2, y2]\u683c\u5f0f\n            x1, y1, w, h = bbox\n            x2, y2 = x1 + w, y1 + h\n            boxes.append([x1, y1, x2, y2])\n            labels.append(ann['category_id'])\n\n        if self.transform:\n            image = self.transform(image)\n\n        return {\n            'image': image,\n            'boxes': np.array(boxes) if boxes else np.zeros((0, 4)),\n            'labels': np.array(labels) if labels else np.zeros(0),\n            'image_id': image_id\n        }\n<\/code><\/pre>\n<h4>2. \u65e0\u76d1\u7763\u57df\u9002\u5e94\u8bad\u7ec3\u793a\u4f8b<\/h4>\n<pre><code class=\"language-python\"># \u4f7f\u7528D-MASTER\u7b97\u6cd5\u8fdb\u884c\u65e0\u76d1\u7763\u57df\u9002\u5e94\u8bad\u7ec3\nfrom d_master import DMasterTrainer\n\n# 1. \u914d\u7f6e\u6e90\u57df\u548c\u76ee\u6807\u57df\u6570\u636e\ncfg = DMasterConfig()\n\n# \u6e90\u57df\u914d\u7f6e\ncfg.SOURCE.DATA_ROOT = &quot;path\/to\/rsna-bsd1k&quot;\ncfg.SOURCE.ANN_FILE = &quot;annotations\/instances_full.json&quot;\ncfg.SOURCE.DOMAIN = &quot;rsna&quot;\n\n# \u76ee\u6807\u57df\u914d\u7f6e\ncfg.TARGET.DATA_ROOT = &quot;path\/to\/inbreast&quot;\ncfg.TARGET.ANN_FILE = &quot;annotations\/inbreast_val.json&quot;\ncfg.TARGET.DOMAIN = &quot;inbreast&quot;  # \u76ee\u6807\u57df\uff1aINBreast\n\n# \u6a21\u578b\u4e0e\u8bad\u7ec3\u914d\u7f6e\ncfg.MODEL.NAME = &quot;d-master&quot;\ncfg.MODEL.BACKBONE = &quot;resnet50&quot;\ncfg.TRAIN.EPOCHS = 20\ncfg.TRAIN.BATCH_SIZE = 2\ncfg.TRAIN.LR = 0.0001\ncfg.EVAL.FPI_THRESHOLDS = [0.3]  # \u91cd\u70b9\u8bc4\u4f300.3 FPI\u4e0b\u7684\u654f\u611f\u6027\n\n# 2. \u542f\u52a8UDA\u8bad\u7ec3\ntrainer = DMasterTrainer(cfg)\ntrainer.train()\n\n# 3. \u8bc4\u4f30\u8de8\u57df\u6027\u80fd\uff08\u76ee\u6807\u57dfINBreast\u4e0a\u7684\u654f\u611f\u6027\uff09\n# \u8bad\u7ec3\u5b8c\u6210\u540e\uff0c\u65e5\u5fd7\u4f1a\u8f93\u51fa\uff1a\n# &quot;Source Domain (RSNA-BSD1K) Sensitivity@0.3FPI: 0.82&quot;\n# &quot;Target Domain (INBreast) Sensitivity@0.3FPI: 0.78&quot;  # \u6bd4\u4f20\u7edfSOTA\u63d0\u534717%\n<\/code><\/pre>\n<p><strong>\u5173\u952e\u8bf4\u660e<\/strong>\uff1aD-MASTER\u7684\u6838\u5fc3\u6539\u8fdb\u662f&#8221;\u63a9\u7801\u9000\u706b\u7279\u5f81\u91cd\u5efa&#8221;\u4e0e&#8221;\u81ea\u9002\u5e94\u7f6e\u4fe1\u5ea6\u4f2a\u6807\u7b7e\u8fc7\u6ee4&#8221;\uff0c\u53ef\u6709\u6548\u63d0\u53d6\u57df\u4e0d\u53d8\u7279\u5f81\uff08\u5982\u80bf\u5757\u7684\u8fb9\u7f18\u3001\u5bc6\u5ea6\u7279\u5f81\uff09\uff0c\u51cf\u5c11\u6e90\u57df\u4e0e\u76ee\u6807\u57df\u7684\u5206\u5e03\u5dee\u5f02\uff1b\u8bad\u7ec3\u65f6\u76ee\u6807\u57df\u4ec5\u8f93\u5165\u56fe\u50cf\uff08\u65e0\u6807\u6ce8\uff09\uff0c\u901a\u8fc7\u4f2a\u6807\u7b7e\u5f15\u5bfc\u6a21\u578b\u5b66\u4e60\uff0c\u65e0\u9700\u624b\u52a8\u6807\u6ce8\u76ee\u6807\u57df\u6570\u636e\uff0c\u5927\u5e45\u964d\u4f4e\u4e34\u5e8a\u5e94\u7528\u6210\u672c\u3002<\/p>\n<h4>3. \u6548\u679c\u53ef\u89c6\u5316\u4e0e\u90e8\u7f72\u5efa\u8bae<\/h4>\n<p><strong>\u6548\u679c\u53ef\u89c6\u5316\uff1a<\/strong><\/p>\n<ol>\n<li><strong>\u8fb9\u754c\u6846\u5bf9\u6bd4\u56fe<\/strong>\uff1a\u5728\u4e73\u817aX\u5149\u7247\u4e0a\u53e0\u52a0&#8221;\u4e13\u5bb6\u6807\u6ce8\u8fb9\u754c\u6846\uff08\u7ea2\u8272\uff09&#8221;\u4e0e&#8221;\u6a21\u578b\u9884\u6d4b\u8fb9\u754c\u6846\uff08\u7eff\u8272\uff09&#8221;\uff0c\u76f4\u89c2\u5c55\u793a\u5c0f\u75c5\u7076\uff08\u5982\u5fae\u9499\u5316\uff09\u7684\u68c0\u6d4b\u7cbe\u5ea6<\/li>\n<li><strong>FPI-\u654f\u611f\u6027\u66f2\u7ebf<\/strong>\uff1a\u7ed8\u5236\u4e0d\u540cFPI\u9608\u503c\uff080.1\/0.3\/0.5\uff09\u4e0b\u7684\u654f\u611f\u6027\u53d8\u5316\u66f2\u7ebf\uff0c\u5bf9\u6bd4\u6a21\u578b\u4e0e\u4f20\u7edfSOTA\u7684\u6027\u80fd\u5dee\u5f02\uff08\u5982D-MASTER vs \u666e\u901aFaster R-CNN\uff09<\/li>\n<li><strong>\u8de8\u57df\u6027\u80fd\u70ed\u529b\u56fe<\/strong>\uff1a\u5c55\u793a\u6a21\u578b\u5728RSNA-BSD1K\uff08\u6e90\u57df\uff09\u3001INBreast\u3001DDSM\uff08\u76ee\u6807\u57df\uff09\u4e0a\u7684\u654f\u611f\u6027\u5206\u5e03\uff0c\u9a8c\u8bc1\u6cdb\u5316\u80fd\u529b<\/li>\n<\/ol>\n<p><strong>\u90e8\u7f72\u5efa\u8bae\uff1a<\/strong><\/p>\n<ol>\n<li><strong>\u7cfb\u7edf\u96c6\u6210<\/strong>\uff1a\u5c06\u8bad\u7ec3\u597d\u7684\u6a21\u578b\u96c6\u6210\u5230\u653e\u5c04\u79d1PACS\uff08\u5f71\u50cf\u5f52\u6863\u548c\u901a\u4fe1\u7cfb\u7edf\uff09\uff0c\u901a\u8fc7API\u63a5\u53e3\u8bfb\u53d6\u4e73\u817aX\u5149\u7247\uff0c\u5b9e\u65f6\u8f93\u51fa\u6076\u6027\u533a\u57df\u8fb9\u754c\u6846\u4e0e\u7f6e\u4fe1\u5ea6\uff0c\u8f85\u52a9\u533b\u751f\u9605\u7247<\/li>\n<li><strong>\u6a21\u578b\u8f7b\u91cf\u5316<\/strong>\uff1a\u82e5\u90e8\u7f72\u5230\u8fb9\u7f18\u8bbe\u5907\uff08\u5982\u79fb\u52a8\u9605\u7247\u7ec8\u7aef\uff09\uff0c\u53ef\u4f7f\u7528TensorRT\u5bf9\u6a21\u578b\u91cf\u5316\uff08FP32\u2192FP16\uff09\uff0c\u6216\u91c7\u7528YOLOv8n\u7b49\u8f7b\u91cf\u7ea7\u6a21\u578b\uff0c\u5c06\u63a8\u7406\u65f6\u95f4\u63a7\u5236\u5728500ms\u4ee5\u5185<\/li>\n<li><strong>\u5408\u89c4\u6027<\/strong>\uff1a\u6570\u636e\u96c6\u867d\u53ef\u7528\u4e8e\u7814\u7a76\uff0c\u4f46\u90e8\u7f72\u65f6\u9700\u786e\u4fdd\u8f93\u5165\u7684\u4e73\u817aX\u5149\u7247\u5df2\u5b8c\u6210\u60a3\u8005\u53bb\u6807\u8bc6\u5316\uff08\u7b26\u5408HIPAA\u3001GDPR\u7b49\u9690\u79c1\u6cd5\u89c4\uff09\uff0c\u4e14\u6a21\u578b\u9700\u901a\u8fc7\u533b\u5b66\u4f26\u7406\u5ba1\u6838\uff0c\u660e\u786e&#8221;\u8f85\u52a9\u8bca\u65ad\u800c\u975e\u66ff\u4ee3\u533b\u751f&#8221;<\/li>\n<\/ol>\n<h3>\uff08\u56db\uff09\u6570\u636e\u96c6\u6837\u4f8b\u5c55\u793a\uff08\u6587\u4ef6\u5939\u7ed3\u6784\u4e0e\u6838\u5fc3\u5143\u7d20\uff09<\/h3>\n<pre><code>rsna-bsd1k\/\n\u251c\u2500\u2500 annotations\/\n\u2502   \u251c\u2500\u2500 instances_full.json    # \u5168\u91cf1000\u5f20\u56fe\u50cf\u7684\u6807\u6ce8\n\u2502   \u2514\u2500\u2500 instances_val.json     # \u9a8c\u8bc1\u96c6\u56fe\u50cf\u6807\u6ce8\n\u2514\u2500\u2500 images\/\n    \u251c\u2500\u2500 train\/                 # \u8bad\u7ec3\u96c6\u56fe\u50cf\uff08\u7ea6800\u5f20\uff09\n    \u2502   \u251c\u2500\u2500 mammo_001.jpg\n    \u2502   \u251c\u2500\u2500 mammo_002.jpg\n    \u2502   \u2514\u2500\u2500 ...\n    \u2514\u2500\u2500 val\/                   # \u9a8c\u8bc1\u96c6\u56fe\u50cf\uff08\u7ea6200\u5f20\uff09\n        \u251c\u2500\u2500 mammo_801.jpg\n        \u251c\u2500\u2500 mammo_802.jpg\n        \u2514\u2500\u2500 ...\n<\/code><\/pre>\n<p><strong>\u6807\u6ce8\u8bf4\u660e<\/strong>\uff1a\u4ee5<code>instances_full.json<\/code>\u4e2d\u7684\u67d0\u6761\u6807\u6ce8\u4e3a\u4f8b\uff0c\u6076\u6027\u75c5\u4f8b\u7684\u6807\u6ce8\u683c\u5f0f\u4e3a\uff1a<\/p>\n<pre><code class=\"language-json\">{\n    &quot;image_id&quot;: 1,\n    &quot;bbox&quot;: [156, 230, 45, 38],\n    &quot;category_id&quot;: 1,\n    &quot;id&quot;: 1001\n}\n<\/code><\/pre>\n<p>\u5176\u4e2d<code>bbox<\/code>\u5bf9\u5e94\u4e73\u817aX\u5149\u7247\u4e2d\u6076\u6027\u80bf\u5757\u7684\u4f4d\u7f6e\uff08xmin=156\uff0cymin=230\uff0cwidth=45\uff0cheight=38\uff0c\u5355\u4f4d\u4e3a\u50cf\u7d20\uff09\uff0c<code>category_id=1<\/code>\u4ee3\u8868&#8221;malignant&#8221;\uff08\u6076\u6027\uff09\uff1b\u826f\u6027\u75c5\u4f8b\u65e0<code>annotations<\/code>\u6761\u76ee\uff0c\u4ec5\u5728<code>images<\/code>\u4e2d\u8bb0\u5f55\u56fe\u50cf\u4fe1\u606f\u3002<\/p>\n<hr \/>\n<h2>\u4e09\u3001\u76f8\u5173\u6570\u636e\u96c6\u5bf9\u6bd4<\/h2>\n<h3>\u4e3b\u6d41\u4e73\u817a\u5f71\u50cf\u6570\u636e\u96c6\u6982\u89c8<\/h3>\n<table>\n<thead>\n<tr>\n<th>\u6570\u636e\u96c6\u540d\u79f0<\/th>\n<th>\u6570\u636e\u89c4\u6a21<\/th>\n<th>\u6807\u6ce8\u7c7b\u578b<\/th>\n<th>\u4e3b\u8981\u7279\u70b9<\/th>\n<th>\u9002\u7528\u573a\u666f<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td><strong>RSNA-BSD1K<\/strong><\/td>\n<td>1000\u5f20X\u5149\u7247<\/td>\n<td>\u8fb9\u754c\u6846\u6807\u6ce8<\/td>\n<td>\u4e13\u5bb6\u7ea7\u6807\u6ce8\uff0c\u8de8\u57df\u7814\u7a76\u652f\u6491<\/td>\n<td>BCDM\u3001UDA\u7814\u7a76<\/td>\n<\/tr>\n<tr>\n<td><strong>CBIS-DDSM<\/strong><\/td>\n<td>10,000+\u5f20X\u5149\u7247<\/td>\n<td>\u8fb9\u754c\u6846+ROI<\/td>\n<td>\u9ad8\u5206\u8fa8\u7387\uff0c\u6807\u51c6\u6807\u6ce8<\/td>\n<td>\u80bf\u5757\u68c0\u6d4b\u4e0e\u5206\u7c7b<\/td>\n<\/tr>\n<tr>\n<td><strong>INbreast<\/strong><\/td>\n<td>410\u5f20X\u5149\u7247<\/td>\n<td>\u8fb9\u754c\u6846+BI-RADS<\/td>\n<td>\u9ad8\u8d28\u91cf\u6807\u6ce8\uff0c\u591a\u89c6\u89d2<\/td>\n<td>\u7b97\u6cd5\u9a8c\u8bc1\u57fa\u51c6<\/td>\n<\/tr>\n<tr>\n<td><strong>BreakHis<\/strong><\/td>\n<td>9,109\u5f20\u75c5\u7406\u5207\u7247<\/td>\n<td>8\u79cd\u4e9a\u578b\u5206\u7c7b<\/td>\n<td>\u7ec6\u7c92\u5ea6\u75c5\u7406\u5206\u7c7b<\/td>\n<td>\u75c5\u7406\u4e9a\u578b\u8bc6\u522b<\/td>\n<\/tr>\n<tr>\n<td><strong>TCGA-BRCA<\/strong><\/td>\n<td>1,000+\u75c5\u4f8b<\/td>\n<td>\u591a\u6a21\u6001\u6570\u636e<\/td>\n<td>\u57fa\u56e0-\u5f71\u50cf-\u4e34\u5e8a\u8054\u5408<\/td>\n<td>\u9884\u540e\u5206\u6790\u7814\u7a76<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<h3>\u6570\u636e\u96c6\u9009\u62e9\u5efa\u8bae<\/h3>\n<pre><code class=\"language-python\"># \u6839\u636e\u7814\u7a76\u9700\u6c42\u9009\u62e9\u5408\u9002\u7684\u6570\u636e\u96c6\ndef select_dataset(research_goal):\n    if research_goal == &quot;\u80bf\u5757\u68c0\u6d4b\u4e0e\u5206\u7c7b&quot;:\n        return [&quot;CBIS-DDSM&quot;, &quot;INbreast&quot;]\n    elif research_goal == &quot;\u75c5\u7406\u4e9a\u578b\u8bc6\u522b&quot;:\n        return [&quot;BreakHis&quot;]\n    elif research_goal == &quot;\u591a\u6a21\u6001\u9884\u540e\u5206\u6790&quot;:\n        return [&quot;TCGA-BRCA&quot;]\n    elif research_goal == &quot;\u8f7b\u91cf\u7ea7\u6a21\u578b\u9a8c\u8bc1&quot;:\n        return [&quot;UCI Wisconsin&quot;]\n    else:\n        return [&quot;RSNA-BSD1K&quot;]  # \u9ed8\u8ba4\u63a8\u8350\n<\/code><\/pre>\n<h3>\u6570\u636e\u4f7f\u7528\u5efa\u8bae<\/h3>\n<table>\n<thead>\n<tr>\n<th>\u9700\u6c42\u573a\u666f<\/th>\n<th>\u63a8\u8350\u6570\u636e\u96c6<\/th>\n<th>\u5173\u952e\u4f18\u52bf<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>\u80bf\u5757\u68c0\u6d4b\u4e0e\u5206\u7c7b<\/td>\n<td>CBIS-DDSM + INbreast<\/td>\n<td>\u9ad8\u5206\u8fa8\u7387+\u6807\u51c6\u6807\u6ce8<\/td>\n<\/tr>\n<tr>\n<td>\u75c5\u7406\u4e9a\u578b\u8bc6\u522b<\/td>\n<td>BreakHis<\/td>\n<td>8\u79cd\u4e9a\u578b\u7ec6\u7c92\u5ea6\u6807\u6ce8<\/td>\n<\/tr>\n<tr>\n<td>\u591a\u6a21\u6001\u9884\u540e\u5206\u6790<\/td>\n<td>TCGA-BRCA<\/td>\n<td>\u57fa\u56e0-\u5f71\u50cf-\u4e34\u5e8a\u8054\u5408\u6570\u636e<\/td>\n<\/tr>\n<tr>\n<td>\u8f7b\u91cf\u7ea7\u6a21\u578b\u9a8c\u8bc1<\/td>\n<td>UCI Wisconsin<\/td>\n<td>\u4f4e\u8ba1\u7b97\u5f00\u9500\uff0c\u5feb\u901f\u539f\u578b\u5f00\u53d1<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<h3>\u6ce8\u610f\u4e8b\u9879<\/h3>\n<ol>\n<li><strong>\u4f26\u7406\u5408\u89c4<\/strong>\uff1a\u90e8\u5206\u6570\u636e\u96c6\u9700\u7b7e\u7f72\u6570\u636e\u4f7f\u7528\u534f\u8bae\uff08\u5982TCGA-BRCA\uff09<\/li>\n<li><strong>\u9884\u5904\u7406<\/strong>\uff1a\u5f71\u50cf\u6570\u636e\u9700\u7edf\u4e00\u5f52\u4e00\u5316\uff08\u5982<code>Pixel \u2208 [0,1]<\/code>\uff09\u5e76\u589e\u5f3a\uff08\u65cb\u8f6c\/\u7ffb\u8f6c\uff09<\/li>\n<li><strong>\u504f\u5dee\u4fee\u6b63<\/strong>\uff1a\u4f7f\u7528Stratified K-fold\u907f\u514d\u7c7b\u522b\u4e0d\u5747\u8861\u5bfc\u81f4\u7684\u8bc4\u4f30\u504f\u5dee<\/li>\n<\/ol>\n<hr \/>\n<h2>\u56db\u3001\u603b\u7ed3<\/h2>\n<p>RSNA-BSD1K\u6570\u636e\u96c6\u4e3a\u4e73\u817a\u764c\u68c0\u6d4b\u548c\u65e0\u76d1\u7763\u57df\u9002\u5e94\u7814\u7a76\u63d0\u4f9b\u4e86\u9ad8\u8d28\u91cf\u7684\u57fa\u51c6\u6570\u636e\uff0c\u5177\u6709\u4ee5\u4e0b\u7279\u70b9\uff1a<\/p>\n<ul>\n<li><strong>\u6743\u5a01\u6807\u6ce8<\/strong>\uff1a200\u4e2a\u6076\u6027\u75c5\u4f8b\u7531\u4e13\u5bb6\u653e\u5c04\u79d1\u533b\u751f\u6807\u6ce8<\/li>\n<li><strong>\u8de8\u57df\u652f\u6491<\/strong>\uff1a\u652f\u6301\u65e0\u76d1\u7763\u57df\u9002\u5e94\u7b97\u6cd5\u5f00\u53d1\u4e0e\u9a8c\u8bc1<\/li>\n<li><strong>\u4efb\u52a1\u4e13\u7528<\/strong>\uff1a\u4e13\u95e8\u9488\u5bf9BCDM\u4efb\u52a1\u8bbe\u8ba1\uff0c\u8986\u76d6\u591a\u5c3a\u5ea6\u75c5\u7076<\/li>\n<li><strong>\u6613\u4e8e\u4f7f\u7528<\/strong>\uff1a\u6807\u51c6COCO\u683c\u5f0f\uff0c\u4e0e\u4e3b\u6d41\u6846\u67b6\u65e0\u7f1d\u5bf9\u63a5<\/li>\n<\/ul>\n<p>\u8be5\u6570\u636e\u96c6\u4e3a\u7814\u7a76\u8005\u63d0\u4f9b\u4e86\u4ece\u7b97\u6cd5\u5f00\u53d1\u5230\u4e34\u5e8a\u90e8\u7f72\u7684\u5b8c\u6574\u6570\u636e\u652f\u6491\uff0c\u662f\u4e73\u817a\u764cAI\u68c0\u6d4b\u9886\u57df\u7684\u91cd\u8981\u8d44\u6e90\u3002<\/p>\n<hr \/>\n<p><strong>\u6570\u636e\u6765\u6e90\uff1a<\/strong> <a href=\"https:\/\/dianshudata.com\/dataDetail\/13749\">\u5178\u67a2\u6570\u636e<\/a><br \/>\n<strong>\u8054\u7cfb\u65b9\u5f0f\uff1a<\/strong> 010-53385803<br \/>\n<strong>\u5546\u4e1a\u5408\u4f5c\uff1a<\/strong> contact@yeez.tech<\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u6700\u65b0\u4e73\u817aX\u5149\u7247\u6570\u636e\u96c6\uff1a\u4e13\u4e3a\u4e73\u817a\u764c\u68c0\u6d4bAI\u6a21\u578b\u4f18\u5316\u8bbe\u8ba1\uff0c\u8986\u76d6\u94bc\u9776\u5f71\u50cf\u7684\u5c0f\u75c5\u7076\u68c0\u6d4b\u4e0e\u65e0\u76d1\u7763\u57df\u9002\u5e94\uff08UDA\uff09\u7814\u7a76\uff0c\u9002\u7528\u4e8eDetectron2\/MMDetection\u6846\u67b6\u7684COCO\u683c\u5f0f<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_seopress_robots_primary_cat":"","_seopress_titles_title":"","_seopress_titles_desc":"","_seopress_robots_index":"","site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"default","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"","footer-sml-layout":"","ast-disable-related-posts":"","theme-transparent-header-meta":"","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"default","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"footnotes":""},"categories":[173],"tags":[223,224,222,95],"class_list":["post-2069","post","type-post","status-publish","format-standard","hentry","category-173","tag-ai","tag-224","tag-222","tag-95"],"_links":{"self":[{"href":"https:\/\/dianshudata.com\/story\/wp-json\/wp\/v2\/posts\/2069","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/dianshudata.com\/story\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/dianshudata.com\/story\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/dianshudata.com\/story\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/dianshudata.com\/story\/wp-json\/wp\/v2\/comments?post=2069"}],"version-history":[{"count":1,"href":"https:\/\/dianshudata.com\/story\/wp-json\/wp\/v2\/posts\/2069\/revisions"}],"predecessor-version":[{"id":2070,"href":"https:\/\/dianshudata.com\/story\/wp-json\/wp\/v2\/posts\/2069\/revisions\/2070"}],"wp:attachment":[{"href":"https:\/\/dianshudata.com\/story\/wp-json\/wp\/v2\/media?parent=2069"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/dianshudata.com\/story\/wp-json\/wp\/v2\/categories?post=2069"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/dianshudata.com\/story\/wp-json\/wp\/v2\/tags?post=2069"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}