Newsroom

Your attention and support are the driving force of Lingzhi's continuous progress

Lingzhi Technology: is face recognition slow and unsafe? Edge algorithm to understand

The development of Internet, AI, IOT and other technologies continuously catalyze the expansion of the data ocean. How to deal with the increasingly huge data flow has become the subject of continuous research.


With the development of science and technology, the innovation of “green than blue” is constantly emerging. According to a cbinsights data, the global IOT market is expected to reach $170 billion in 2019, which means that the data to be processed will grow wildly. The surge of data processing demand not only promotes the development of cloud computing, but also stimulates the concept of edge computing. At present, Amazon, Microsoft, Google and other giants have their own edge computing layout. Emerging technology enterprises in this field have also begun to emerge. Lingzhi technology is a new star in this field.


Shanghai Lingzhi technology, established in July 2016, is an AI enabled enterprise. At present, its product line mainly focuses on smart life scenarios, and other application scenarios include the automotive industry. According to Zhang Kejun, general manager of Lingzhi technology, Professor Su Jianbo, the founder of the company, is an authoritative figure in the field of intelligent robots in China, with decades of industry accumulation in this field. Therefore, it has rich experience in technology development and product transformation. Lingzhi technology has more than 20 invention patents, and obtained ten million level pre-A round financing led by Taihao capital last year.


Lingzhi’s products include smart home robots and face recognition locks. “We focus on building a safe community in our life, providing a complete set of solutions from door lock, access control, gate machine and community monitoring, including corresponding modules of face recognition, emotion recognition, object recognition and gesture recognition.” Zhang Kejun said. In fact, lingzhi technology has started AI research since 1998, and has completed the research on 5000 person face recognition algorithm, 7 emotion recognition algorithm, 10 intention recognition algorithm and 30 object recognition algorithm. It is worth mentioning that the AI technology of computer vision recognition is developed based on the independent edge algorithm, dexterous edge algorithm (laf2.0) owned by Lingzhi technology.


As described at the beginning of the article, edge computing is gaining attention. Last year, “edge computing” appeared for the first time in Gartner’s “top 10 strategic technology trends.”. So, what is edge computing? Baidu Encyclopedia’s explanation is: it refers to an open platform that integrates network, computing, storage and application core capabilities at the side close to the object or data source to provide the nearest end service nearby.


Zhang Kejun explained these professional terms more vividly. “Edge algorithm is a branch of deep learning algorithm.” He said. Its characteristics are: it not only has the efficient recognition ability of traditional deep learning algorithm, but also avoids the problems such as large resource occupation and difficulty in landing. “That is, what used to be downloaded and done on a large server can now be done on a local embedded platform.” He said that edge computing has strong real-time performance, high cost performance, less environmental requirements, and can run without network.


The AI technology of computer vision recognition based on self-developed edge algorithm has become the technical advantage of Lingzhi technology, which also distinguishes its products from most visual products on the market.


Take the door lock and access control products of face recognition as an example. The operation logic of the general face recognition door lock product is: the user takes a picture before the door lock, and then the photo is sent to the cloud for recognition through the network, and then the recognition result is fed back to the door lock, and then the door is opened. In fact, this process takes some time. Moreover, in some poor network environment, the process of traditional algorithm will be slower. “At the same time, there are hidden dangers in the process of face recognition after the collection.” Zhang added.


Through the edge algorithm, lingzhi technology products can identify the process locally, which saves the time of back and forth transmission, improves the efficiency, but also improves the security, and realizes the low-cost implementation of complex algorithms.


Although the current market is the “world” of cloud computing, Zhang Kejun believes in the future development potential of edge computing. “Edge computing is a big development direction in the future.” He believes that edge computing can be used in many scenarios, such as the automotive industry. As we all know, the information update of the car needs every second. However, if the mobile phone signal is not good and the network environment is poor, the information update will be delayed, thus causing certain security risks.
Of course, lingzhi technology has built its own unique advantages in technology and application. In terms of product landing, it also has its own story. “Lingzhi technology aims to provide products and traditional industries with artificial intelligence wings. Therefore, we have accumulated experience in different industries in the process of providing grounded gas and landing enabling solutions for many industries. ” His experience in these industries is very valuable. In addition, he added that the service awareness and quick response ability of the whole team of Lingzhi technology are also characteristics of the enterprise.