皮膚癌的ai診斷

1、由一篇Nature引發(fā)的思考,PPT制作、主講 朱明哲文獻查找、翻譯、Word文檔制作 李卓,,,是什么研究成果震撼了生物醫(yī)學界？,生物醫(yī)學傳感領域的新突破,,百度搜索引擎都膜拜的第一作者,榮獲斯坦福大學AI領域的理學博士并修德克薩斯大學奧斯汀分校電子計算機工程及純數(shù)學雙專業(yè),Dermatologist-level classification of skin canc

2、erwith deep neural networks,利用深度神經(jīng)網(wǎng)絡，進行皮膚科專家級別的皮膚癌診斷,這項研究成果被x-mol官網(wǎng)評為2017年2月全球科學技術(shù)十大突破之五,這種所謂的深度神經(jīng)網(wǎng)絡就是近年來經(jīng)常提及 但又充滿神秘的,人工智能 (Artificial Intelligence),,,人工智能我們該如何理解？,從數(shù)學上看,深度神經(jīng)網(wǎng)絡(Deep Neural Networks),,卷積神經(jīng)網(wǎng)絡(Convolu

3、tional Neural Networks),淺析CNNs,Logistic回歸模型,淺析CNNs,卷積的處理看起來使得問題復雜了，這是否會增大計算難度，降低效率？,課堂討論環(huán)節(jié)顯然讓課堂效率提高了而它正是卷積過程的體現(xiàn),Layer L1在座近100位同學都提出了自己的觀點,Layer L2通過小組討論，初步得出問題的解答,Layer L3對每個小組的討論結(jié)果進行評估，考慮其對問題的偏離程度（權(quán)重），得出最終結(jié)果,CNNs的

4、工作就是將全局映射轉(zhuǎn)化為區(qū)域映射的過程。,另一個現(xiàn)實中的例子是圖像處理,,,卷積神經(jīng)網(wǎng)絡在該課題中如何運用？,Classical diagnosing method,Initial clinical screening,Dermoscopic analysis,Biopsy,Histopathological examination,Observing undermicroscope,Diagnosticreport,For

5、mation of AI dermatologist ——A process of machine learning,Figure 1 | Deep CNN layout. Formed by Google Inception v3 CNN architecture,Step1 They demonstrated classification of skin lesions using a single CNN, tra

6、ined end-to-end from images directly, using only pixels and disease labels as inputs.,Figure 2 | A schematic illustration of the taxonomy and example test set images,a, A subset of the top of the tree-structured taxonomy

7、 of skin disease.b, Malignant and benign example images from two disease classes.,Step 2 They trained a CNN using a dataset of 129,450 clinical images, consisting of 2,032 different diseases.,Step 3 Grasp the characteri

8、stic message of pictures by analyzing pixels.,Figure 3 | Skin cancer classification performance of dermatologists (a) and CNN (b).,Practical diagnose ——Analyzing up-loaded picture from mobile devices,Fig

9、ure 4 | t-SNE visualization of the last hidden layer representations in the CNN for four disease classes,Real case analysis as support information,,,AI診斷的發(fā)展前景如何？,Outfitted with deep neural networks, mobile devices can po

10、tentially extend the reach of dermatologists outside of the clinic. It is projected that 6.3 billion smartphone subscriptions will exist by the year 2021 and can therefore potentially provide low-cost universal a

11、ccess to vital diagnostic care.,——Ericsson Mobility Report, 2016,,,除了診斷病癥AI還給我們帶來了什么？,Deep learning algorithms, powered by advances in computation and very large datasets, have recently been shown to exceed human perfor

12、mance in visual tasks such as playing Atari games, strategic board games like Go and object recognition.,AI應用于生活的幾個例子,2017年5月23日~5月27日 “第二次圍棋人機大戰(zhàn)”,論決策和價值評估，人類真的輸了！,Demis Hassabis，創(chuàng)業(yè)公司DeepMind創(chuàng)始人。,“AI是人類探索世界的工具?！?,,屬于我們自己