|
|
<p style="white-space: normal;"><img class="rich_pages js_insertlocalimg" data-backh="253" data-backw="578" data-ratio="0.43859649122807015" data-s="300,640" src="http://mmbiz.qpic.cn/mmbiz_gif/dx3KibwtDhvd30KriaSbvhoncZUg9ZbETjscR1rN5dSCfznIuBFJJWmiauFOFKJq6LKLDom0CYRBzWtBO9Qp0jHhw/640?wx_fmt=gif" data-type="gif" data-w="798" style="text-align: center;width: 100%;height: auto;" /><br /></p><p style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", " ingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;">乳腺癌是世界女性健康的第一杀手，而我国是乳腺癌发病率增长速度最快的国家之一，并且以每年2%的速度递增。当人类持续多年锲而不舍地尝试开发和综合利用各种最先进的技术手段以及药物治疗之后发现大部分晚期癌症患者的预后仍然像一个谜时。人类终于不得不接受把癌症的治疗端口前移的选择，集中在早诊断、早发现、早预防、早治疗的维度上。因此，针对乳腺癌的威胁，国家自2009年起就发起了关于两癌筛查的国家免费公共卫生项目。</span></p><section data-support="96编辑器" data-style-id="22631" style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "ingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);box-sizing: border-box !important;overflow-wrap: break-word !important;"><section style="margin-bottom: 10px;max-width: 100%;text-align: center;line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><section style="padding-right: 5px;padding-left: 5px;max-width: 100%;box-sizing: border-box;border-width: 6px;border-bottom-style: solid;border-color: rgb(7, 126, 154);display: inline-block;overflow-wrap: break-word !important;"><p style="max-width: 100%;min-height: 1em;box-sizing: border-box !important;overflow-wrap: break-word !important;"><strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">01</strong></p></section><section style="padding-right: 10px;padding-left: 10px;max-width: 100%;box-sizing: border-box;border-width: 6px;border-bottom-style: solid;border-color: rgb(102, 182, 195);display: inline-block;overflow-wrap: break-word !important;"><p style="max-width: 100%;min-height: 1em;box-sizing: border-box !important;overflow-wrap: break-word !important;"><strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">生理差异 使我国不能“照抄”欧美</strong></p></section></section></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "ingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;">虽然国家发起了免费两癌筛查的公益项目，但由于大众的医学健康意识尚未系统建立、技术尚不成熟、人力成本等因素，筛查进度不尽人意。目前对于乳腺癌的早期筛查，最可靠的方式是医学影像，发达国家多采用钼靶作为常规检查方式，美国的家庭医生会建议40~49岁的女性每2年钼靶摄影筛查1次，大于50岁的女性建议根据个人情况进行钼靶摄影筛查，而建议50-74岁女性每2年钼靶摄影筛查1次。中国也在2018年发布的《中国乳腺癌筛查与早诊早治指南》中，明确了在乳腺癌筛查时，采用钼靶结合超声的方式。</span></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "ingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);text-align: center;line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><img class="rich_pages" data-ratio="1" data-s="300,640" data-type="jpeg" data-w="474" src="http://mmbiz.qpic.cn/mmbiz_jpg/bfCedyicff1Ucom0vKpMSFtJsbnVNSI4MVZiaIAEIrtXibPNLaNIWzgaKphJgUErkDDEt8iasGAVDmEtMs39QZrM1g/640?wx_fmt=jpeg" style="box-sizing: border-box !important;overflow-wrap: break-word !important;width: 100%;visibility: visible !important;height: auto;" data-backw="474" data-backh="474" /></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "ingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;letter-spacing: 1.5px;box-sizing: border-box !important;overflow-wrap: break-word !important;">由于亚洲女性生理结构与欧美人种的差异，</span><strong style="max-width: 100%;letter-spacing: 1.5px;box-sizing: border-box !important;overflow-wrap: break-word !important;">亚洲人钼靶摄影的影像辨识早期乳腺癌的难度远大于欧美</strong><span style="max-width: 100%;letter-spacing: 1.5px;box-sizing: border-box !important;overflow-wrap: break-word !important;">，而</span>超声筛查乳腺癌敏感性低，特异性差，只作为钼靶的辅助检查手段，但在亚太地区，超声作为主要筛查方式或许更加符合国情。但是，超声属于动态图像检查，对于检查医生的水平、集中度都有着较高的要求，因此以超声作为主要的乳腺癌筛查手段同样存在困难。</span></section><section data-support="96编辑器" data-style-id="22631" style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "ingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);box-sizing: border-box !important;overflow-wrap: break-word !important;"><section style="margin-bottom: 10px;max-width: 100%;text-align: center;line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><section style="padding-right: 5px;padding-left: 5px;max-width: 100%;box-sizing: border-box;border-width: 6px;border-bottom-style: solid;border-color: rgb(7, 126, 154);display: inline-block;overflow-wrap: break-word !important;"><p style="max-width: 100%;min-height: 1em;box-sizing: border-box !important;overflow-wrap: break-word !important;"><strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">02</strong></p></section><section style="padding-right: 10px;padding-left: 10px;max-width: 100%;box-sizing: border-box;border-width: 6px;border-bottom-style: solid;border-color: rgb(102, 182, 195);display: inline-block;overflow-wrap: break-word !important;"><p style="max-width: 100%;min-height: 1em;box-sizing: border-box !important;overflow-wrap: break-word !important;"><strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">超声破局 技术难度大</strong></p></section></section></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "ingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;">既然超声筛查存在弊端，那么我们是否可以用一些方案来补充呢？显然，AI系统能够更好的解放医生，降低医生的工作强度，而依托于大数据，也能使检查结果更加准确，但采用一直以来的截取图像，再用AI分析的传统AI分析方式，对于超声检查而言意义不大。</span></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "ingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;">AI技术在医学影像辅助诊断方向上的发展，随着<strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">算法、数据成熟程度</strong>等因素的不断进化。<strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">最初AI主要面向2D影像</strong>，因为算法和数据的公开程度分为两个阶段，第一阶段多以肺结节/眼底等拥有公开数据集的产品，第二阶段是以乳腺钼靶等病种为代表的无公开数据集的病种。自此之后，面向3D影像的AI辅助诊断产品被开发出来，广泛应用于CT/MR等需要进行三维重建的领域，尽管大大增加了AI分析的难度，但在目前的应用当中，效果卓绝。</span></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "ingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;">然而AI识别仍然只停留在静态图像的解析上，面对超声这种一系列连续的视频信号，此前并没有AI产品能够做到，因为这一领域，<strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">既没有公开的数据集，也没有现成的算法模型</strong><strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">，甚至连数据怎么采集、怎么标注都还未有公认的方法</strong>。因为二维图像相对容易获得且处理算法难度小，因此全国以及全球范围内，对超声问题的探索大多都还着眼于二维影像，但对于实际应用场景来讲，二维图病灶信息不够充分，分析的准确性存在天然的瓶颈，距真正的临床辅助诊断有很大差距。这就意味着，<strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">超声AI辅助诊断产品不仅要对视频数据的采集、存储和识别设计新的方法，还需要设计一套专门的算法</strong>。</span></section><section data-support="96编辑器" data-style-id="22631" style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "ingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);box-sizing: border-box !important;overflow-wrap: break-word !important;"><section style="margin-bottom: 10px;max-width: 100%;text-align: center;line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><section style="padding-right: 5px;padding-left: 5px;max-width: 100%;box-sizing: border-box;border-width: 6px;border-bottom-style: solid;border-color: rgb(7, 126, 154);display: inline-block;overflow-wrap: break-word !important;"><p style="max-width: 100%;min-height: 1em;box-sizing: border-box !important;overflow-wrap: break-word !important;"><strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">03</strong></p></section><section style="padding-right: 10px;padding-left: 10px;max-width: 100%;box-sizing: border-box;border-width: 6px;border-bottom-style: solid;border-color: rgb(102, 182, 195);display: inline-block;overflow-wrap: break-word !important;"><p style="max-width: 100%;min-height: 1em;box-sizing: border-box !important;overflow-wrap: break-word !important;"><strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">时效性与准确率矛盾</strong></p></section></section></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;">超声研发目前面临的最主要的技术难题就是对实时性以及准确率的双高要求，也就是既要达到实时性，同时又要有极高的准确率。CT/MR的图像采集和诊断是分开来进行，因此AI系统有充足的时间来对图像进行处理，但由于超声的诊断流程及特点，其要求超声AI<strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">必须达到实时辅助诊断的效果</strong>。AI需在医师操作超声仪器时同时查看扫查的视频影像进行实时诊断，并给出诊断结果，再将结果反馈给医师，才能有效的辅助医师进行诊断。</span></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;">实时分析对AI的算法和算力提出了更高的要求，尤其是在医疗领域，对准确率异常高的要求，而这就意味着模型的复杂度将会大大增加，以便提取更加丰富的特征，而超声AI又不能牺牲时间来换取准确率，因此算法方面面临巨大挑战。同时，由于视频数据量巨大，所以日常超声检查并不会大量存储视频数据，通常是只留几张具有代表性的图片，给超声数据的采集带来极大挑战。</span></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;">另外，优秀的模型还需要依赖高质量的标注数据，而超声视频数据，每个病人通常十分钟左右，以<strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">30帧/秒</strong>来算，那么一个病人的二维超声影像就有<strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">10*60*30=18000张</strong>，这些都需要具有丰富超声医学知识的专家来标注。而我国目前的现状是超声医师本身就存在近十万的缺口，其日常工作就非常繁忙，要找到专业的人员完成如此艰巨的标注任务其挑战难度可以想象。</span></section><section data-support="96编辑器" data-style-id="22631" style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);box-sizing: border-box !important;overflow-wrap: break-word !important;"><section style="margin-bottom: 10px;max-width: 100%;text-align: center;line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><section style="padding-right: 5px;padding-left: 5px;max-width: 100%;box-sizing: border-box;border-width: 6px;border-bottom-style: solid;border-color: rgb(7, 126, 154);display: inline-block;overflow-wrap: break-word !important;"><p style="max-width: 100%;min-height: 1em;box-sizing: border-box !important;overflow-wrap: break-word !important;"><strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">04</strong></p></section><section style="padding-right: 10px;padding-left: 10px;max-width: 100%;box-sizing: border-box;border-width: 6px;border-bottom-style: solid;border-color: rgb(102, 182, 195);display: inline-block;overflow-wrap: break-word !important;"><p style="max-width: 100%;min-height: 1em;box-sizing: border-box !important;overflow-wrap: break-word !important;"><strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">创造新模型 医准智能乳腺超声智能检测系统</strong></p></section></section></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;">作为超声检测的一个具体部位，乳腺自身的特点，也加剧了AI的技术研发难度。主要表现在两个方面。<strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">一是乳腺病灶跟脂肪的区分困难；二是乳腺病灶良恶性判断困难。</strong></span></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;">很多脂肪切面与病灶相似，因此超声检查时需对疑似病灶多角度扫查，降低假阳性诊断，而对于视频检测算法，由于实时性的需求，它更多的是关注当前接收到的影像，对当前接收的影像进行分析。而通常的算法只能识别并分析病灶部分，而无法识别周边信息，无法像真正的医生一样多角度综合判断。所以，乳腺超声视频检测和分类算法的研发之路颇为曲折。</span></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;">医准智能的研发团队克服了以上的种种难题，发布了<strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">国内首个乳腺超声智能检测系统</strong>，医准智能为这套系统建立了新的模型，针对乳腺超声的两个技术难点做出了针对性的调整。在《中国乳腺癌筛查与早诊早治指南》发布后，医准智能就着手研发乳腺癌筛查AI，2019年率先开发出医准智能乳腺钼靶智能检测系统，2020年，医准智能又完成了AI识别2D图像到动态图像的跨越，攻克了超声AI的高地。</span></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);text-align: center;line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><img class="rich_pages" data-ratio="0.5003872966692486" data-s="300,640" data-type="png" data-w="1291" src="http://mmbiz.qpic.cn/mmbiz_png/bfCedyicff1Ucom0vKpMSFtJsbnVNSI4MN4kBk4XoicECVPF6E2f8chHD3RuxUS1OBCDS3vquTNbdypwTfEd8vUw/640?wx_fmt=png" style="box-sizing: border-box !important;overflow-wrap: break-word !important;width: 100%;visibility: visible !important;height: auto;" data-backw="578" data-backh="289" /></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><br /></section><ul class="list-paddingleft-2" style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);width: 577.422px;overflow-wrap: break-word !important;"><li style="max-width: 100%;font-size: 14px;color: rgb(136, 136, 136);box-sizing: border-box !important;overflow-wrap: break-word !important;"><section style="max-width: 100%;line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;">针对由于乳腺病灶跟脂肪的区分困难而带来假阳性增加的问题：在FPN利用多尺度特征的基础上，加入了时间维度信息，并采用3D CNN，同时提取病灶时间维度和空间维度的三维特征，综合以上特征，有效的剔除了脂肪等假阳性。</section></li><li style="max-width: 100%;font-size: 14px;color: rgb(136, 136, 136);box-sizing: border-box !important;overflow-wrap: break-word !important;"><section style="max-width: 100%;line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;">针对病灶良恶性判断困难问题：为了获得更加准确的良恶性分类结果，首先，通过病灶各个截面的视频还原出病灶的三维结构；其次，加入注意力机制，将周边组织的特征以不同权重加入到病灶特征中，充分利用了病灶以及其周边信息，获得了更加准确的分类效果。</section><section style="max-width: 100%;line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><br /></section></li></ul><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;">医准智能乳腺超声智能检测系统<strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">不需要对设备改动，且不需要调整现有工作流程</strong>，在医生扫查病人的同时，<strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">AI服务器进行实时分析，并在界面进行标记提示</strong>，对于仅毫秒级闪现的病灶也能精准抓取，能够有效的避免医生由于视觉疲劳以及视觉敏感度不够而产生的漏诊。</span></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);text-align: center;line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><img class="rich_pages __bg_gif" data-ratio="0.5625" data-type="gif" data-w="640" src="http://mmbiz.qpic.cn/mmbiz_gif/bfCedyicff1VCVNHpqwiaENND8y8ibu8CCebIlkKsdmLWj8F5rs28yicWyCjW2EPPib08qEXKMuia5ibpOh1T5Umysx0g/640?wx_fmt=gif" style="box-sizing: border-box !important;overflow-wrap: break-word !important;width: 100%;visibility: visible !important;height: auto;" data-backw="578" data-backh="325" /></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);text-align: center;line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 12px;color: rgb(136, 136, 136);box-sizing: border-box !important;overflow-wrap: break-word !important;">自动病灶BI-RADS分类</span></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;">作为应用于临床的系统，医准智能乳腺超声智能检测系统还具有如下五大亮点：</span></section><ul class="list-paddingleft-2" style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);width: 577.422px;overflow-wrap: break-word !important;"><li style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><section style="max-width: 100%;line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;">计算速度快，低延时；</section></li></ul><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;"> 系统采用神经网络架构搜索方法（NAS），使用RTX2080Ti。达到每秒处理速度＞ 50帧，且检测结果延迟＜0.09 秒，对于仅毫秒级闪现的病灶也能精准抓取。</span></section><ul class="list-paddingleft-2" style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);width: 577.422px;overflow-wrap: break-word !important;"><li style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><section style="max-width: 100%;line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;">病灶识别检出率高；</section></li></ul><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;">通过卷积神经网络特征融合模拟恢复3D，从而得出病灶性质良恶性判断。</span></section><ul class="list-paddingleft-2" style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);width: 577.422px;overflow-wrap: break-word !important;"><li style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><section style="max-width: 100%;line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;">假阳性（误报）低；</section></li></ul><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;">系统通过在所有影像中筛选有意义帧进行判断，从而有效降低假阳性率。</span></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);text-align: center;line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><img class="rich_pages __bg_gif" data-ratio="0.5421875" data-type="gif" data-w="640" src="http://mmbiz.qpic.cn/mmbiz_gif/bfCedyicff1Ucom0vKpMSFtJsbnVNSI4MVEY0UxyF4Fh3iaGkU6kANAl4iawTQ4TIbgjkLsX6J2c3VkIrjUahy2Qg/640?wx_fmt=gif" style="box-sizing: border-box !important;overflow-wrap: break-word !important;width: 100%;visibility: visible !important;height: auto;" data-backw="578" data-backh="314" /></section><ul class="list-paddingleft-2" style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);width: 577.422px;overflow-wrap: break-word !important;"><li style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><section style="max-width: 100%;line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;">智能分割 自动测量；</section></li></ul><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;">相对于仅仅利用病灶的某个截面特征进行属性分析，该系统还可以对包含病灶所有信息的整段视频进行分析，充分利用病灶的各个截面信息，对病灶整体进行更加充分的属性分析，同时给出病灶最大截面以及长短径、面积等定量信息。</span></section><ul class="list-paddingleft-2" style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);width: 577.422px;overflow-wrap: break-word !important;"><li style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><section style="max-width: 100%;line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;">结构化报告生成</section></li></ul><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;">一键获取，加快影像报告诊断效率；智能化超声所见及诊断，协助临床获取疾病详细文本数。</span></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">结语：</strong></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;">医准智能开发的乳腺超声智能检测系统，或许将彻底改变乳腺癌的筛查，甚至改变整个AI辅助诊断行业的格局。以往的AI产品多依托于影像科，使用DR、CT、MR等影像设备，可绝大多数影像科设备无法将检查环境前移，而超声设备在此方面有着得天独厚的优势，体型小巧无辐射的超声可以脱离科室的束缚，将检查环境前移，让乳腺癌筛查更加便捷。</span></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;">此外，过去的AI辅助诊断产品全部都是采用截取静态图像，AI进行分析的操作，而<strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">AI直接分析动态图像</strong>的医准智能乳腺超声智能检测系统<strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">完全颠覆了AI辅助诊断领域的固有模式</strong>，为医疗AI解锁全新技能。</span></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;">从该产品早期合作医院的使用反馈来看，超声医师使用医准智能乳腺超声智能检测系统时不必改变操作方式，AI系统即时可给出参考意见并自动标注，为超声医师的判断提供强有力的支持，不仅缩短医师判断时间内，还能减少医师的精力消耗，且不再需要再搭配专门的录入人员。同时做到大幅度提升乳腺筛查的准确率、大幅降低超声医师的工作强度并为医院节省人力成本，因此产品广受超声科医生喜爱。</span></section><section style="max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;white-space: normal;background-color: rgb(255, 255, 255);line-height: 2em;margin-top: 5px;box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 15px;box-sizing: border-box !important;overflow-wrap: break-word !important;">最后，据悉医准智能正在探索AI系统+影像设备集成模块化的可能，未来，颠覆式的AI系统加上模块化的影像产品势必将覆盖更多病种的检查。中国的两癌筛查已经做了10年，但10年间乳腺癌的死亡人数仍然在不断上升，很大一部分原因就是发现的太晚，乳腺超声智能检测系统能够助力乳腺癌在筛查阶段更早的发现病患，降低乳腺癌死亡率。</span></section><section style="white-space: normal;text-align: center;"><span style="font-size: 12px;background-color: rgb(255, 255, 255);font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;widows: 1;">-end-</span><br /></section><p style="font-variant-numeric: normal;font-variant-east-asian: normal;white-space: normal;max-width: 100%;min-height: 1em;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;widows: 1;font-size: 16px;text-align: center;background-color: rgb(255, 255, 255);box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 12px;box-sizing: border-box !important;overflow-wrap: break-word !important;">部分图片来源于网络，如有侵权，请联系小编进行删除</span></p><p style="font-variant-numeric: normal;font-variant-east-asian: normal;white-space: normal;max-width: 100%;min-height: 1em;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;widows: 1;font-size: 16px;text-align: center;background-color: rgb(255, 255, 255);box-sizing: border-box !important;overflow-wrap: break-word !important;"><span style="max-width: 100%;font-size: 12px;box-sizing: border-box !important;overflow-wrap: break-word !important;">以上文章及作品皆为仁成所有，转载请注明出处</span></p><p style="font-variant-numeric: normal;font-variant-east-asian: normal;white-space: normal;max-width: 100%;min-height: 1em;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;widows: 1;font-size: 16px;text-align: center;background-color: rgb(255, 255, 255);box-sizing: border-box !important;overflow-wrap: break-word !important;"><br /></p><section data-role="outer" label="Powered by gulangu" style="font-variant-numeric: normal;font-variant-east-asian: normal;white-space: normal;max-width: 100%;font-family: -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;widows: 1;font-size: 16px;background-color: rgb(255, 255, 255);box-sizing: border-box !important;overflow-wrap: break-word !important;"><section data-tools="gulangu" data-id="92886" style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;"><section style="padding: 10px;max-width: 100%;box-sizing: border-box;color: rgb(46, 164, 166);overflow-wrap: break-word !important;"><section style="max-width: 100%;display: flex;justify-content: center;align-items: center;box-sizing: border-box !important;overflow-wrap: break-word !important;"><section style="max-width: 100%;font-size: 12px;line-height: 1.6em;text-align: center;box-sizing: border-box !important;overflow-wrap: break-word !important;">▼</section><section data-brushtype="text" style="padding-right: 10px;padding-left: 10px;max-width: 100%;box-sizing: border-box;line-height: 1.6em;letter-spacing: 1px;text-align: center;overflow-wrap: break-word !important;"><strong style="max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;">往期精彩回顾</strong></section><section style="max-width: 100%;font-size: 12px;line-height: 1.6em;text-align: center;box-sizing: border-box !important;overflow-wrap: break-word !important;">▼</section></section></section></section></section><p style="white-space: normal;text-align: center;"><a target="_blank" href="http://mp.weixin.qq.com/s?__biz=MjM5Nzk4MTg5Mw==&mid=2654232621&idx=1&sn=39131b363631ef5d9e1c65e8101e873e&chksm=bd117b878a66f2913ba785eca9d471d104d6113bbc5941161253beb12566b603cea80325727e&scene=21#wechat_redirect" textvalue="你已选中了添加链接的内容" data-itemshowtype="0" tab="innerlink" data-linktype="1"><span class="js_jump_icon h5_image_link" data-positionback="static" style="top: auto;left: auto;right: auto;bottom: auto;"><img class="rich_pages" data-backh="160" data-backw="578" data-ratio="0.2765625" data-s="300,640" src="http://mmbiz.qpic.cn/mmbiz_jpg/dx3KibwtDhvdzocwrib8S1Kj9ticfkuIbcsyXzzXb7pSCJPAxd8ibqBibicNJ6mvSGbbCzQNDnqYdTjdzhklfsARvFRg/640?wx_fmt=jpeg" data-type="jpeg" data-w="1280" style="width: 100%;top: auto;left: auto;right: auto;bottom: auto;height: auto;" /></span></a></p><p style="white-space: normal;text-align: center;"><a target="_blank" href="http://mp.weixin.qq.com/s?__biz=MjM5Nzk4MTg5Mw==&mid=2654233252&idx=2&sn=2bccbb19ad9870008157f8fe32b3ec60&chksm=bd11790e8a66f018fe863d87f068450cc96914181a71b334ce7d96ccb8096a110a56613fa0bf&scene=21#wechat_redirect" textvalue="你已选中了添加链接的内容" data-itemshowtype="0" tab="innerlink" data-linktype="1"><span class="js_jump_icon h5_image_link" data-positionback="static" style="top: auto;left: auto;margin: 0px;right: auto;bottom: auto;"><img class="rich_pages js_insertlocalimg" data-backh="160" data-backw="578" data-cropselx1="0" data-cropselx2="578" data-cropsely1="0" data-cropsely2="160" data-ratio="0.27685185185185185" data-s="300,640" src="http://mmbiz.qpic.cn/mmbiz_jpg/dx3KibwtDhvdk36fSTBu3DKSPKfJvRQreZPhKuocOwIeDRSk2valkiabg7lqFfavyKbBWuwtCLBiaLV1UJxrxBFbQ/640?wx_fmt=jpeg" data-type="jpeg" data-w="1080" style="width: 100%;top: auto;left: auto;right: auto;bottom: auto;height: auto;margin: 0px;" /></span></a></p><p style="white-space: normal;text-align: center;"><a target="_blank" href="http://mp.weixin.qq.com/s?__biz=MjM5Nzk4MTg5Mw==&mid=2654232566&idx=2&sn=528956df2663b209eb3791f5cfc40811&chksm=bd11445c8a66cd4ae8382adcaaf81ac4a8ca7bcaa890fa953fafc519808c42edc9c3d2c68a57&scene=21#wechat_redirect" textvalue="你已选中了添加链接的内容" data-itemshowtype="0" tab="innerlink" data-linktype="1"><span class="js_jump_icon h5_image_link" data-positionback="static" style="top: auto;left: auto;right: auto;bottom: auto;"><img class="rich_pages js_insertlocalimg" data-backh="160" data-backw="578" data-ratio="0.2765625" data-s="300,640" src="http://mmbiz.qpic.cn/mmbiz_jpg/dx3KibwtDhvdaoLyYB5XBZDWW5tA6THRqUVY3Cxd5PsjL5gTIW3LNx8AM84EfEcJkKryUqgBAXHriaiaYQcUJyP5w/640?wx_fmt=jpeg" data-type="jpeg" data-w="1280" style="width: 100%;height: auto;" /></span></a></p><p style="white-space: normal;text-align: center;"><img class="rich_pages js_insertlocalimg" data-backh="269" data-backw="578" data-cropselx1="0" data-cropselx2="578" data-cropsely1="0" data-cropsely2="270" data-ratio="0.46707739699653444" data-s="300,640" src="http://mmbiz.qpic.cn/mmbiz_png/dx3KibwtDhvdaoLyYB5XBZDWW5tA6THRqTufQJgss2BJSGLrET00ibjaCkGv7BX6SRF9golHNV9d2gn07eWvvmIA/640?wx_fmt=png" data-type="png" data-w="2597" style="width: 100%;height: auto;" /></p>
|
|
发表于 2020-10-19 22:38:43