谷歌人工智能學(xué)會診斷眼疾

Google’s DeepMind has developed artificial intelligence to diagnose diseases by analysing medical images, in what could be the first significant application of AI in healthcare.

谷歌(Google)的DeepMind已開發(fā)出通過分析醫(yī)學(xué)影像診斷疾病的人工智能，這可能是人工智能在醫(yī)療領(lǐng)域的首個重大應(yīng)用。

London-based DeepMind has crunched data from thousands of retinal scans to train an AI algorithm to detect signs of eye disease more quickly and efficiently than human specialists.

總部位于倫敦的DeepMind處理了數(shù)千張視網(wǎng)膜掃描圖的數(shù)據(jù)，以訓(xùn)練一種人工智能算法探測眼病跡象——比人類專家更迅速和更高效率。

Findings have been submitted to a medical journal by the company after “promising signs” from a two-year partnership with the National Health Service and London’s Moorfields Eye Hospital, one of the best-known in the world for ophthalmology. The technology could enter clinical trials in a few years if results pass a peer review by academics.

該公司已將研究結(jié)果遞交一家醫(yī)學(xué)刊物，此前該公司與英國國民醫(yī)療服務(wù)體系(NHS)以及全球最知名的眼科醫(yī)院之一、倫敦的Moorfields Eye Hospital為期兩年的合作出現(xiàn)“頗有希望的進(jìn)展”。如果研究結(jié)果通過學(xué)者們的同行評議，這項技術(shù)有望在幾年后投入臨床試驗。

“In areas like medical imaging, you can see we’re going to make really tremendous progress in the next couple of years with artificial intelligence,” Dominic King, clinical lead for DeepMind Health, told the Financial Times. DeepMind Health

臨床主管多米尼克•金(Dominic King)告訴英國《金融時報》：“在醫(yī)學(xué)影像等領(lǐng)域，你可以看到未來兩年我們將借助人工智能取得真正重大的進(jìn)展。”

DeepMind’s algorithm has been trained using anonymised 3D retinal scans provided by Moorfields and painstakingly labelled for signs of disease by doctors. The DeepMind division has now begun discussing clinical trials with hospitals including Moorfields.

DeepMind的算法利用Moorfields提供、并已由醫(yī)生煞費苦心地標(biāo)出疾病跡象的匿名3D視網(wǎng)膜掃描圖進(jìn)行訓(xùn)練。DeepMind團(tuán)隊現(xiàn)在已開始與包括Moorfields在內(nèi)的醫(yī)院討論臨床試驗。

Because the images provide rich data with millions of pixels of information, the algorithm can learn to analyse them for signs of three of the biggest eye diseases: glaucoma, diabetic retinopathy and age-related macular degeneration.

由于這些影像提供了數(shù)以百萬計像素信息的豐富數(shù)據(jù)，算法可以學(xué)會對這些數(shù)據(jù)進(jìn)行分析，以尋找三大眼疾的跡象：青光眼、糖尿病視網(wǎng)膜病變和老年性黃斑退化。

Peng Tee Khaw, director of research and development at Moorfields, said: “I am optimistic that what we learn from this research will benefit people around the world and help put an end to avoidable sight loss.”

Moorfields研發(fā)總監(jiān)Peng Tee Khaw表示：“我樂觀地相信，這項研究帶給我們的收獲將造福于世界各地的人們，幫助杜絕可以避免的失明。”

The artificial intelligence is “generalised”, according to Dr King, meaning it can be applied to other kinds of images. DeepMind said the next stage would involve training the algorithm to analyse radiotherapy scans, through a partnership with University College London Hospitals, and mammograms, in partnership with Imperial College London.

DeepMind的金博士表示，這種人工智能是“普遍化的”，這意味著它可應(yīng)用于其他類型的影像。DeepMind表示，下一步將訓(xùn)練算法分析放療掃描圖(通過與University College London Hospitals合作)以及乳腺X光照片(與倫敦帝國理工學(xué)院合作)。