In many sci-fi blockbusters, the protagonist can control the screen of devices such as computers through gestures, and even directly control objects. Similar scenarios, in fact, do not need to wait for the unreachable future, it is happening now within reach.

This remote control technology is called "sensory control" in the industry, and the earliest somatosensory control product that entered the public's field of vision is Microsoft's Xbox One host, which can support voice commands and Kinect gesture operations.

Kinect gesture function of Microsoft Xbox

It is precisely because of seeing Xbox One that Liu Jiale, who was still working on Kingsoft in 2014, was very excited. He believed that in the future of human-computer interaction, somatosensory control will play an important role, so he founded Yelang Intelligence.

Although from 2014 to 2019, the process of the somatosensory control industry has not been smooth. In November 2015, Microsoft removed the Kinect gesture function of the main interface of Xbox One; another product that supports touch screen and voice control and is also more familiar to domestic users. Lao Luo ’s TNT workstation soon faded from the topic center. Until no one cares.

However, Ye Langzhi, who has been focusing on somatosensory control, has gone from dormant to the foreground. In July 2017, Ye Lang Intelligent received millions of angel rounds of financing and began formal operations; in the first half of this year, the company's first product for ordinary enthusiasts, Hello X, launched JD crowdfunding, which soon exceeded the target amount of fundraising. For the future of somatosensory control, Ye Langzhi even boldly predicted, "Reform traditional human-computer interaction within 5 years!"

Of course, in addition to optimism, the team also believes that we need to standardize sensible control technology and improve interaction stability.

With curiosity about this "black technology" and expectations for the future of somatosensory control, Huazhuan.com interviewed Liu Jiale, founder and CEO of Yelang Intelligence. The following is the full text of the interview.

What is a better way of human-computer interaction?

Huazhuan.com: Ye Lang Intelligence has been doing surface virtualization technology since 2014. Is this considered relatively early in the industry?

Liu: It was very early in the industry, because there were very few companies that did this kind of surface virtualization or somatosensory control. When we first started doing it in 2014, only the Xbox game consoles released by Microsoft in the world involved this technology, and artificial intelligence was not as hot as it was today.

Huazhuan.com: Why did you want to start your business from this angle?

Liu: I worked at Kingsoft before 2014. I saw Microsoft products and thought that there might be more opportunities in this field in the future.

The reason why I chose to do human-computer interaction is because I feel that traditional methods like mouse and keyboard are gradually unable to serve us well. Now the devices are getting smaller and smaller, from tablets, mobile phones to watches, may become projection-based in the future. We are wondering if there can be some better ways of interaction to adapt to the changes in equipment. And we feel that there are not too many companies digging in this direction, and we will have some opportunities to overtake in corners.

Huazhuan.com: There are many ways to transform traditional human-computer interaction, and the current voice input solutions seem to occupy the mainstream. But why do you still choose somatosensory control instead of voice input?

Liu: Voice input is not very suitable in many scenes. If you need to perform voice interaction with the device, the recognition rate will be very low if the distance is relatively long. In addition, in the office environment, it is not so convenient to always use voice to interact with the device.

For example, Luo Yonghao's TNT workstation realizes human-device interaction through touch screen and voice. There are some problems with its voice interaction. The first is that there is a problem with privacy protection during the interaction; the second aspect is that its voice cannot be modified; the third is that the user needs to organize the language first when speaking, to ensure that it can be Device identification.

Let the machine understand people

Huazhuan.com: Can you briefly introduce the principle of somatosensory control technology?

Liu: The principle is very simple, that is to make the machine understand people.

Its process is probably: when the camera sees you, the machine will make a judgment on your behavior. If the camera sees you coming, the machine will automatically light up the screen; if you can see your hand, you can control the machine with your hand, that is, let the machine understand some of your behavior. The camera first shoots, and then recognizes whether it is a human through an integrated algorithm, and whether there is a hand to control the machine

Huazhuan.com: Compared with TNT workstations or other touch screen devices, what are the advantages of Ye Lang's solution?

Liu: Compared with the touch screen method of TNT workstation, our products can leave the device and do gesture control. By making a gesture in one direction, you can directly control the device, this interaction is closer to the natural state of the person.

In fact, the work we are doing is not intended to completely replace the traditional way of interaction. Gesture control is still a blue ocean, and no standards have been formed. The direction we are currently working on is to define rules. When there is no standard yet, we want to use somatosensory interaction with traditional interaction.

For example, the Hello X product we released, when you are away from the computer, you can control the device with gestures; when you are closer, you can use the touch screen; when you need to type in the office scene, you can also use the traditional keyboard. The environment we hope to create is that when I want to control a device, I don't need to care about what control is better for this thing. Put your hand on the screen, you can touch, no matter whether this screen supports, this is the combination with the traditional touch screen, is also the way we are trying.

Huazhuan.com: So, do you think you need to use a variety of interaction methods first, and then slowly form user education and user habits?

Liu: Yes, this is an inevitable process.

The mouse and keyboard have been working together for decades, and the two have not been turned into one thing because they are responsible for different things. In fact, the two can't be put together, we have actually developed a product demo that puts the mouse and keyboard together. Although this method cannot be popularized at present, it must be a direction.

Once hard problems such as battery life are resolved, a product may explode. The second situation is that if technology such as 3D display breaks out, it is up to us at this time. Because after using 3D display, it is impossible to use keyboard and mouse for interaction. This may not happen within a year or two, but it must be a trend. Before that, we need to have a certain technical foundation. It is impossible to start research after the new trend is fully formed.

Huazhuan.com: Do you need guidance and training in the early stage of the somatosensory control operation gestures?

Liu: At present, our operation gestures are actually only raising hands and gliding. This is why we call this product Hello X. Hello means you need to raise your hand, sayHello with the device, and then you can interact by sliding. For first-time users, there will be operation guidance on our software interface. I think proficiency is not difficult for users.

Focus on expanding the B-end market

Huazhuan.com: Yelang Intelligent currently mainly adopts ToB or ToC sales model?

Liu: Our current positioning is based on B-end customers. There will only be one product on the C side, Hello X, which will carry a lot of functions, and there may be many variants.

Huazhuan.com: How did B-end do it? What customers or scenes have demand for Yelang's products now?

Liu: In some industry scenarios where we naturally have a combination, such as projection and television, we focus on layout. There are some other areas that need to be cooperated, such as education, training, medical treatment, etc., and we can do it in scenarios that are related to people.

For example, TV, there are many ways of cooperation. The way we expect is integrated, which means that the TV will bring us such a function as soon as it leaves the factory. In this way, we can better control whether it is shipment or accuracy. Of course, you can also sell it with the TV as an accessory like Hello X.

Huazhuan.com: What is the price of Hello X? What is the current sales volume?

Liu: The price is 399 yuan. We are still crowdfunding on JD.com, and we reached the crowdfunding target of RMB 100,000 within a week of launch, and we can start shipping recently.

Huazhuan.com: Ye Lang's current focus is on technology research and development, or application promotion and user education?

Liu: Our core is still in the research and development of the underlying technology and algorithms, which is time-consuming and is also the direction we pay more attention to. We will also do things like application promotion, because this is a bridge, but our focus is still on algorithms.

Huazhuan.com: What is the technical strength of Yelang Intelligence?

Liu: We are a leader in the subdivision of specific gesture interaction methods and methods in China. Because there are no particularly large giants in this regard, and there is no unicorn in the market, it is still a blue ocean. From the perspective of the underlying technology, the companies are not much different, but we are doing it in different directions, so we will go forward in the direction of gestures or interaction.

We are now a startup company, and we ca n’t compare technology with industry giants. What we can do better than them is to subdivide the interaction to make users feel more comfortable. It will involve a lot of unique logic, as well as very unique scenarios, not to say that the algorithm is very powerful, you can do a good job of subdividing the field. We have some advantages in this respect.

Huazhuan.com: How many people are there in the team?

Liu: Our company currently has about 10 people. At the beginning of the establishment of the Guangzhou company in 2014, there were very few people. The team was gradually expanded after the establishment of the Zhongshan company in 2017.

Huazhuan.com: Where does Yelang Smart currently have an office location, and what are the considerations?

Liu: There are currently two offices in Guangzhou and Zhongshan. The R & D team is mainly in Guangzhou, and the production and sales will be in Zhongshan. After all, the home appliance market in Zhongshan is still relatively prosperous and has a certain influence throughout the country.

Huazhuan.com: Is the company seeking investment and in what areas will the funds be used?

Liu: The financing of Pre-A is already in progress. The funds are mainly used in the R & D team, and the market will also be deployed.

The main purpose of our financing is to obtain funds. In addition to the funds, we also hope to obtain some industry resources and have some opportunities for cooperation with head companies. This will not only allow us to get orders, but also get a deeper understanding of industry information. In terms of talents, we also hope to obtain some resources.

In the next 5 years, a directional market

Huazhuan.com: The company has proposed to reform the traditional human-computer interaction method within 5 years. Are you still optimistic about this expectation?

Liu: This statement was made when we completed the last round of financing.

I think it is not possible to reform traditional human-computer interaction methods at present. On the one hand there are problems with standardization, on the other hand there are technical reasons.

So how do we see the market or user needs for new interactive methods? In fact, there are many devices, which are difficult to operate in the traditional way. For example, projection, the screen is projected on the wall or the desktop, the traditional interaction method has actually failed. Projection has been getting better and better in recent years, and there is a need for interaction. We are meeting this kind of demand. From this point, our direction is definitely no problem. In addition, as more and more electronic devices become more and more complex, users will face management problems. How to do it without user management? This is actually the direction of our interaction.

We say that the most ideal interaction is no interaction. For example, as soon as you walk into the hall, the lights are turned on automatically, and as soon as you approach the TV, the TV turns on. This kind of thing may be achieved by some devices such as infrared sensors. Our advantage lies in the accurate judgment of AI on behavior, reducing the possibility of misjudgment.

An important direction we are currently working on is the "virtual workbench", which is based on Hello X to expand. We will make your desk into a full-space interactive closed loop, whether it is on the display or in a three-dimensional space, you can have a very strong interaction.

I think this is a directional market, not to say what level will be reached in five years, this is not under our control. We can make some estimates about this market, but more importantly, we will do our own thing well, no matter what the environment is, we can find a suitable way of survival.