韓貴來 李海霞

摘? 要: 為了更好地適應人工智能時代發(fā)展需求,在醫(yī)學圖像處理教學過程中融入人工智能內(nèi)容。通過對傳統(tǒng)的醫(yī)學圖像處理教學內(nèi)容知識進行重構(gòu),以項目開發(fā)的形式進行醫(yī)學圖像處理教學改革。實踐表明,融合人工智能技術(shù)的醫(yī)學圖像處理教學改革,更能充分調(diào)動起學生學習的積極性和主動性,使學生真正做到學有所用,學以致用。
關(guān)鍵詞: 醫(yī)學圖像處理; 人工智能; 教學改革; 項目開發(fā)
中圖分類號:G642.0? ? ? ? ? 文獻標識碼:A? ? ?文章編號:1006-8228(2020)11-117-03
Abstract: In order to better adapt to the needs of the development in the era of artificial intelligence, the content of artificial intelligence is integrated into the teaching process of medical image processing. Through the reconstruction of the traditional teaching content knowledge of medical image processing, the teaching reform of medical image processing is carried out in the form of project development. The practice shows that the teaching reform of medical image processing integrating artificial intelligence technology can fully arouse the enthusiasm and initiative of students, truly achieve that what they learned is useful and their learning is for the practice.
Key words: medical image processing; artificial intelligence; teaching reform; project development
0 引言
人工智能技術(shù)快速發(fā)展,已經(jīng)從研究領(lǐng)域進入到應用層面。醫(yī)學人工智能作為一個重要的細分領(lǐng)域, 正逐步在醫(yī)療領(lǐng)域發(fā)揮重要作用。2017年7月,國務院印發(fā)了《新一代人工智能發(fā)展規(guī)劃》,提出要發(fā)展智慧醫(yī)療,實現(xiàn)醫(yī)學影像識別、病理分型和智能診斷等。到2020年,我國醫(yī)學影像市場規(guī)模將達到7000億元。隨著采集的醫(yī)學影像數(shù)據(jù)量越來越大,如今醫(yī)院往往缺乏足夠的讀片人員,因此,對于能夠快速處理影像數(shù)據(jù)的人工智能技術(shù)需求十分巨大。人工智能用于影像處理具有多項優(yōu)勢。
首先,人工智能可以大批量快速地處理影像數(shù)據(jù)。只要計算能力夠大,人工智能便可以一次性處理大量影像數(shù)據(jù)。更重要的是,人工智能不會感到疲勞,只要機器正常,可以連續(xù)24小時不停地工作。而醫(yī)生工作時間過長就會產(chǎn)生疲勞,從而增加漏診和誤診的概率。
其次,人工智能判斷更加客觀、準確。微軟2014年提出的ResNet 深度學習網(wǎng)絡(luò)最深高達152 層,利用殘差的思想將ImageNet大規(guī)模圖像分類競賽的Top-5錯誤率降到了3.57%[1]。……