Detection of pulmonary nodules through Deep Learning

Lung cancer has been the leading cause of cancer death in Hong Kong. At an early stage, lung cancer mostly exists in the form of small pulmonary nodules, which appear on medical images as shades of small lumps. Currently, doctors depend on chest CT scans to reveal those nodules. However, each scan often results in hundreds of images. Assuming that going through each image requires 3 seconds, an analysis of these images by the naked eye will take 5 minutes to complete. Such examinations are time consuming, and must rely on the doctors’ experience and sharpness of focus. When Prof. Heng and his team apply deep learning technology to CT scans, they are able to locate the pulmonary nodules in 30 seconds, with an accuracy of 90%. 

The technology, which the CUHK team started working on five years ago, is at the forefront of international medical technology. With positive feedback from the medical sector, Professor Heng expects it to be widely adopted in the next couple of years. ‘Deep learning makes use of advanced training to improve the sensitivity of the technology, so that it is able to tackle a major challenge that a naked-eye examination faces - that is, removing noise and reducing false positives,’ said Professor Heng. He went on to disclose that, in order to further improve the technology, the team would be working with top hospitals in Beijing, to provide solid evidence in support of early diagnosis and treatment of lung cancer. 

Automated Detection of Metastatic Breast Cancer in Histology Images 

Since 1990, the number of breast cancer patients in Hong Kong has been consistently on the rise. It is the most prevalent cancer amongst local women, and the third amongst all cancers. To determine whether a patient has the cancer, doctors often must extract and examine live tissue samples. Using mammograms or MR scans to locate the lump, samples are extracted and examined under the microscope to see if there are signs of tumour and whether the tumour is benign or malignant. A digital histology is of high resolution, often up to one gigabyte in file size - equivalent to a 90-minute high resolution movie. Examining such an image requires a lot of time and energy. 

To solve the problem, the CUHK team has developed a novel deep cascaded convolutional neural network to process the histopathological images. Making use of a fully convolutional network, the model can efficiently and accurately detect the metastatic cancer with a high-resolution score-map. The whole automated analysis process takes about 5~10 minutes, as compared to the 15~30 minutes that are required if examined by the naked eye. In terms of accuracy, the system has achieved a rate of 98.75%, 2% higher than analysis conducted by experienced doctors. This indicates that it is an invaluable reference for clinical diagnosis on breast cancer. 

A key advantage of artificially intelligent deep learning is that it is able to analyse large quantaties of parameters. The more the data, the higher its accuracy. When this automated screening and analysis system is applied to the medical sector, it acts as a tireless assistant to the doctor that can quickly identify the source of an illness, enabling a timely and appropriate treatment.

HKIA serves over 70.5 million travelers in 2016 and handles over 1,100 flights daily. Around 13,000 baggage trolleys are distributed throughout the airport to cope with the enormous passenger flow, and maintaining a steady supply of trolleys has been one of the major concerns of HKIA.  The CUHK team led by Prof. Cheng Chun-hung, Department of Systems Engineering and Engineering Management has collaborated with HKIA in estimating the baggage trolley availability at various trolley racks or pick-up points through machine learning techniques, image-based technologies and existing surveillance CCTV cameras, since 2014.  The Real-time Trolley Supply Monitoring System, which can be connected to iOS and Android apps, enables frontline service providers and management to monitor trolley availability at all pick-up points.  Different signals represent different levels of availability: yellow alert for quantity dropping down to 50 or below, red alert for empty racks and green alert for normal supply of more than 50.  The system makes resources allocation more flexible and efficient, and thereby improves the passenger’s experience. 

Low installation cost, high service efficiency

The CUHK team has put a lot of effort into enhancing the ease of system deployment and reducing the challenges that may arise.  Prof. Cheng Chun Hung said, ‘By using machine learning techniques such as support vector machines and neural networks, we are able to use video data to build trolley detectors that can identify four different types of trolley and conduct the smart counting. Thanks to HKIA for providing the testing site and solid support, the detection rate of the system has increased from 87% to 92%.’ 

‘A camera network infrastructure normally incur high installation cost which covers building, cabling and engineering works.  To minimize costs, our team has devised a data network with edge processing capability to eliminate the need of high bandwidth video transfer while at the same time, maintaining the image quality and data quantity,’ said Prof. Cheng Chun Hung. 

Mr. Chris AuYoung, General Manager, Smart Airport of the AA said, ‘HKIA strives to develop into a smart airport.  Through the application of intelligent data and automation technologies, we hope HKIA can become more efficient, convenient, and enhance the airport experience for our passengers.  This is a good example illustrating our successful collaboration with the local research and development institutions and universities.  This baggage trolley tracking system not only greatly reduces the need for manual checking of the trolleys, but also helps our service provider to replenish the trolleys at specific locations at a timely manner.  As a result, the service level of trolley availability for passengers in the baggage reclaim hall has been improved, which helps us provide a pleasant airport experience to our worldwide passengers.;  He added, ‘Since trolley management is a common challenge for most of the airports in the world, there would be a good opportunity to export such a solution and Hong Kong International Airport would be a good showcase to illustrate the business benefits’, he added. 

Mr. Simon Wong, Chief Executive Officer of the LSCM R&D Centre, said, ‘We are delighted that academics and industry can utilize the funding support from the Innovation and Technology Fund for enabling research, so that technology is deployed more commonly and effectively for our daily living. We hope that this technology will be extended to other areas for a wider application.’ 

Facilitating long term planning and ensuring security

‘The research project has facilitated the resource and service management at HKIA because the frontline staff can now perform timely trolley replenishment. In addition, the system is able to support data-oriented analysis and even big data analysis for a range of long term resource planning as it continuously collects operational data on trolley usage and replenishment’, said Dorbin Ng, a member of the research team.

Tim Chan, also a member of the research team said, ‘Currently, 18 video cameras have been placed in the Baggage Reclaim Hall for monitoring trolley availability. Regarding passenger privacy issue, we have repeatedly worked with HKIA to define the most appropriate video monitoring areas. The system is also able to automatically blur visual contents other than the trolley racks.  All images are encrypted and requires specific client applications to decrypt for viewing.  We want to minimize both the privacy and security concerns.’ 

About the Department of Systems Engineering and Engineering Management

Founded in 1991, the Department is the first of its kind among tertiary institutions in Hong Kong. It is committed to combining technology with management, in a mission to educate a new generation of technologically skilled and managerially adept engineers who can take on new challenges. Its undergraduate programme offers four specialization streams including business information systems; financial engineering; logistics and supply chain management; and service engineering and management. For more information, please visit www.se.cuhk.edu.hk.