Why are we called "recursion"
"Recursion" (recursion) is an algorithm that is widely used in programming languages. It has two major characteristics, one is to call itself, and the other is to simplify the complex. Isn't that the case with the excellent technical people among us? They lead by example, use practical actions to influence those around them, standardize the process, improve efficiency, and benefit the entire team. This is the original intention of our "recursion" column, to record the extraordinary of ordinary Tencent technical people.
There are more than 255 million elderly people over 60 years old in China, accounting for 18% of the total population. Aging and elderly care issues have gradually become a topic of concern. According to incomplete statistics, at least 40 million elderly people experience falls every year, and falls have become the main cause of death from accidental injuries for people over 65 years of age. In many first-tier cities, young people work hard outside during the day. How can we discover the safety risks around the elderly in time, protect the elderly at home, and us who will grow old in the future?
In Tencent, there is a group of people who are paying attention to the elderly. They are focusing on the safety of the elderly, making technology a kind of guardian and companionship, and allowing "AI" to accompany the family. Today, let us walk into this team and listen to a story about "love".
Liu Shuisheng, senior engineer of Tencent Security Platform Department, head of Tencent Instant AI, has in-depth research in the field of AI applications and AI security technology, and has extensive technical experience.
Sijia Tu , senior product manager of Tencent Security Platform Department, senior product manager of Tencent Instant Vision, has long been engaged in the exploration and commercialization of AI-related products, has rich experience in industry products and projects, and continues to focus on "technology for good" technology, products and operations path.
Tang Yanping , Senior Product Manager of Tencent Instant Vision, focuses on the incubation of commercial products in the direction of AI and security. He has more than ten years of product planning and operation experience and rich experience in industry products and projects. Love public welfare, love to share, used to be the head of the Tencent Volunteer Association's green safety network course, and now serves as an intermediate lecturer of Tencent Academy, giving lectures on "PPT Production Strategy", "User Demand Analysis" and other courses.
Wang Nan , the head of the Smart Elderly Care Project Team, is mainly responsible for the preparation, opening, and management team recruitment of Shenzhen Elderly Care Homes. Through the dean responsibility system under the leadership of the council, he actively promotes the reform and exploration of new types of institutions. Build an IoT platform based on Tencent Cloud architecture to provide a series of solutions including monitoring, fire fighting, wrestling, nursing, and rehabilitation.
Unpreventable fall accident for the elderly
Before developing the "Smart Elderly Care Monitoring System" project, we conducted a survey within the team. More than 30% of our colleagues' immediate family members had accidents where the elderly fell at home or outside. If you extend the scope of the survey to friends around you, this data will reach 60%, and some of the elders of colleagues will have some irreversible accidents.
For this reason, we interviewed the China Senior Citizens Foundation and learned that the first five minutes for the elderly to fall are the "golden life-saving five minutes". If the elderly are found to have fallen and take first aid measures within these five minutes, there is an 80% probability that their lives can be saved. This also made us determined to do this project and felt that it was a very meaningful work.
Sample problem: frequent but few records
The biggest difficulty in this project is the lack of data samples. Elderly falls are actually a relatively frequent event, but because of the sudden nature of falls and the concealment of places, it is difficult to record. We searched for data samples and still did not find them, so we tried to solve them by ourselves.
Through the opportunity of Tencent Cloud WeChat, we found a nursing home in Shenzhen and connected to their camera video. But really deep into the actual environment, we found that the training and experimental environment is too much different than usual, such as the angle and light of the camera, the interference of multiple people in the picture, these problems will have an impact on the AI algorithm, which requires us to have Large and diverse samples.
"Falled" products
To this end, everyone in the team joined the team of "Tencent Trial Falling Officials" to imitate the fall of the elderly in various situations in a realistic environment. In order to better simulate the environment where the elderly are alone, they started a fall test in an unmanned environment. Everyone has tried to use various angles, various postures, various speeds, and various lengths of fall to test.
Throwing forward, sideways, face down, and fancy throws, without much consideration, I fell hundreds of times, and fell every corner of the nursing home and laboratory, just like this, and fell the product. "Fall" out.
New problems, new challenges
After version 1.0 was deployed in nursing homes, we were highly recognized by the hospital. However, in the process of product iteration, we also found that in addition to public corridors, the scenes of old people falling are more likely to occur in private spaces such as toilets.
We installed this instant vision product by means of a camera, but the privacy space cannot have a camera, we had a whim, refer to the somatosensory games such as Switch and X-Box, and help us through infrared and other means Guard the elderly in the privacy space. After continuous testing and verification, we finally found a chip called "ToF", which can also detect the fall of the elderly under the premise of protecting privacy.
Development towards home care
In fast-growing high-tech cities like Shenzhen, many young people go to work during the day. There are only elderly people at home. The distance between the company and the family may not be that far, but for the elderly who may fall, it is a race against time and space. To this end, we actively responded to the call of the national policy, including home care for the elderly, and expanded this product from "nursing home" to "home care".
If the elderly at home are willing, we will deploy the equipment to the home environment. Our AI solution can discover some of their safety hazards, and even determine the diseases that the elderly may suffer from changes in gait. The elderly can stay at home freely and take into account safety protection. They can notify their families, communities, and hospitals as soon as possible, so that children working outside will also feel at ease.
Technology with temperature, warmer love
At present, the number of elderly people over 60 in China has reached 255 million, accounting for about 18% of the total population, so elderly care is a top priority in China.
Our parents are getting old, and we will get old in the future. Now we can stand here and use science and technology to promote some more useful things, to promote a country that needs to be paid attention to, and truly achieve "technology Xiangshan", I feel very honored.
The first season of "Recursion" is over. In the ten-episode documentary, what do you see, who do you like the most, and how has your impression of programmers changed? Maybe you still want to know more stories about Tencent technical people? Leave a message and let us know! We will select five lucky readers and give a small gift~
Director: Lu Qianyu
Guests: Liu Shuisheng, Tu Sijia, Tang Yanping, Wang Nan
Producer: Sun Jianxiang, Xiao Baiwei
Later guidance: paulinesong, kylinlu
Special thanks: Tencent Instant Vision team
Welcome to follow our video number: Tencent programmer
Latest video: Tencent’s 22nd anniversary, draw penguins with code
[ Tencent programmer video account exchange salon event preview]
Saturday, November 21 at 2 p.m.
Shenzhen Tencent Binhai Building
If you want to participate, you can add WeChat: journeylife1900
(Remarks: Tencent Technology)