According to the VR network editor, Gesqua Technology, a startup company that implements the gesture recognition module in the TOF program in China, announced that it has completed more than 20 million A rounds of financing, and the investor is Shanghai Yucai Capital. G-fish technology CEO Fang Wenxin said that this round of financing will be mainly used for the development and production of in-car gesture recognition modules.
Previously, Geshui's gesture recognition module was mainly aimed at the AR/VR field and the smart home field. Most of the customers were downstream B-side hardware manufacturers. Fang Wenxin stated that although AR/VR is undoubtedly the future development direction, at this stage, it still has the problems of imperfect hardware product technology, weak demand of C-end users, and low willingness of manufacturers to pay. The smart home manufacturers tend to use voice interaction or Super APP control devices because the indoor environment is relatively simple.
Compared to the above two application industries, the demand for gesture recognition modules in the automotive front loading market is gradually increasing. And the gesture recognition function will be applied to mid-to-high-end models. The depot will pay more attention to the experience and the sensitivity to price is relatively low.
However, whether or not the gesture interaction in the vehicle is really just a need? The current interactive system in the car has undergone three generations of programs, namely, a generation of physical buttons, a second-generation touch screen, and three generations of voice command interactions. In addition to Delphi Delphi equipped with a gesture recognition system for the BMW 7 Series, this cool program has also attracted the attention of many Internet depots and traditional yards, and gestures are gradually becoming the fourth interactive mode in the car.
The representative solutions for in-car voice interaction include out-of-the-way interrogation, car radish, etc. They emphasize that voice is the most natural and familiar interaction method for users and that the solution and industry chain are mature. However, the limitations of voice interaction in the car scene can not be ignored, the first is the accuracy of speech semantic recognition has yet to be improved, many instructions can not be accurately used by the user speech description, and the noisy environment in the car will further reduce the ability to understand instructions.
Fang Wenxin believes that the future interaction in the car must be a combination of multiple technologies, rather than an "otherwise" model. Based on the actual needs of the depot, the gesture recognition solution provider has to do is how to better user experience. The user experience mainly focuses on the speed and accuracy of gesture recognition.
The first problem faced by Extreme Fish Technology is to simplify the algorithm. Before in the entertainment industry, its gesture recognition algorithms all ran on mobile smart device chips. In the interior scene, the algorithm is running on the car chip, and the performance of the car chip is much lower than that of the mobile smart device. Fang Wenxin stated that they simplified and optimized the algorithm and reduced it by two orders of magnitude.
Now Geshui's gesture recognition program can identify 5-10 dynamic and static gestures. The depot can choose 3-5 gestures and customize the corresponding functions. The recognition accuracy is 1-2cm, and the recognition accuracy can reach 95% or more. The delay rate can be controlled within 25ms.
The other is more communication and long cycle. Due to the different product designs and software agreements of each depot, Extreme Fish Technology needs to provide customized solutions for users, with an average of 3-6 months per project cycle. However, Fang Wenxin stated that in the future they will develop universal plug-and-play modules to reduce the marginal cost of development.
In addition, the modules used in the car need to meet vehicle-level standards and can withstand high and low temperatures, light resistance, and corrosion resistance. And in the process of engineering, Extreme Fish Technology needs to control the module's volume, power consumption, weight and other parameters. Fang Wenxin also stated that compared with the binocular scheme, the TOF scheme has strong ability to resist ambient light interference, but at this stage, the industrial chain is not sound enough. It also requires continuous process iterations in the production process.
Extreme Fish Technology now offers three system solutions: central control gesture interaction, HUD gesture interaction, and in-vehicle passenger entertainment interaction. After completing the development of the in-car gesture recognition module, it is imperative for G-fish to rapidly expand its depot customers. Fang Wenxin stated that they have had contact with Internet car companies, traditional car companies and some HUD and digital instrument manufacturers. It also obtained a full-year order from an Internet depot and expects to ship Q1 next year, with an average monthly order volume of around one thousand units. At the same time, Extreme Fish Technology has also become a global innovation and technology partner of a well-known German car company, and will jointly release innovation results in Beijing in mid-December 2017. Next year, Gefish Technology will further expand 1-2 depot customers. It is expected that sales for 2018 will reach 30 million yuan in the whole year.
In terms of market competition, Leap Motion, uSens and Micro Motion all use a binocular solution. As mentioned above, the binocular scheme is susceptible to light interference and is not very suitable for interior scenes. Structured light solutions represented by RealSense and Opzo Midori cannot be used at close range. Fang Wenxin said that gesture recognition based on TOF technology is still not very mature in the market. Gutfish Technology has performed three iterations on both hardware and algorithms. It is a comprehensive solution that integrates algorithm + software + hardware + optics + vehicle gauge, and is expected to have a leading period of at least 9 months.
Gugefish Technology will open a multi-million dollar A+ round of financing at the end of the year, mainly to expand production lines and expand production capacity from 500-1000 units per month to about 10,000 per month.
Floating Board-to-Board Connectors
Floating Board-To-Board Connectors,Side Plug Horizontal Docking Connector,Domestic Board-To-Board Connectors,2.0Mm Board To Board Connector
Dongguan SOLEPIN Electronics Co., Ltd , https://www.wentae.com