{"id":52835,"date":"2019-11-17T00:00:00","date_gmt":"2019-11-16T21:00:00","guid":{"rendered":"https:\/\/prohoster.info\/blog\/blog_prohoster\/preimushhestva-oblachnogo-raspoznavaniya-lits"},"modified":"2020-02-18T14:00:38","modified_gmt":"2020-02-18T11:00:38","slug":"preimushhestva-oblachnogo-raspoznavaniya-lits","status":"publish","type":"post","link":"https:\/\/prohoster.info\/en\/blog\/administrirovanie\/preimushhestva-oblachnogo-raspoznavaniya-lits","title":{"rendered":"Advantages of Cloud Face Recognition","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"

\"Advantages
\nThe Near Future<\/i><\/p>\n

There are several methods by which face recognition systems operate, but generally, it refers to technology capable of identifying a person based on a digital image or a frame from a video source.<\/p>\n

Many smartphone owners use face recognition daily, but in mobile devices, the speed of recognition is not critical, and the number of users is rarely more than one or two. Different technologies are used for office and outdoor systems (in mass recognition). <\/p>\n

Recently discussed on Habr a news item<\/a><\/noindex>: Moscow-based coffee chains 'Pravda Coffee' and OneBucksCoffee have begun testing a face recognition service in their establishments. <\/p>\n

The coffee shops use our technical solution. Today, we will tell you more about it. Of course, we have already talked about the technology itself, but something new has emerged\u2014the solution has become truly cloud-based. And that changes everything.
\n<\/a><\/noindex><\/p>\n

How face recognition technology works<\/h2>\n

\nThe first thing the system must do is to isolate the face in the frame and, using algorithms, verify that it is indeed a human face. <\/p>\n

After the initial detection, various individual features are determined based on fixed points\u2014for example, the distance between the eyes and dozens of other parameters are taken into account. <\/p>\n

Next, other algorithms search various pre-created databases and provide a similarity percentage with the sought data sample. If the similarity percentage is high enough, the face is considered recognized.<\/p>\n

Without going into details (the photo for analysis also needs to be normalized before passing it to the neural network, which reads some descriptor), the main complexity of the solution at the moment lies not in the technologies (algorithms) themselves, but in the implementation.<\/p>\n

Recognition systems are evolving in several directions, classified according to their approach to information processing. Sometimes it is difficult to choose which system will handle a specific task better. <\/p>\n

A Variety of Systems<\/h2>\n

\n\"Advantages
\n
\nData can be processed in the cloud, on local servers deployed within the enterprise security perimeter, or directly on the cameras. <\/p>\n

In the latter case, all analysis is performed by the camera itself, and processed information is sent to the server. The main advantage of the system is its high accuracy and the ability to connect a large number of devices to a single server.<\/p>\n

Despite the apparent simplicity and ease of scaling, this technology also has its downsides. One of them is the high cost. Additionally, there is currently no single standard for the information transmitted to the server by specialized cameras. The data sets can vary significantly among different vendors.<\/p>\n

\"Advantages
\nThe 'simple' face recognition system from Panasonic <\/a><\/noindex><\/i><\/p>\n

IP camera systems with built-in video analytics are less popular than server-based solutions. However, even when using a traditional system based on a recorder and\/or local server, cost savings are not achievable. <\/p>\n

Programs and prices* Face Recognition <\/h2>\n

\n*According to information from open sources.
\n<\/i>
\nConsidering the complexity of algorithms and the high cost of server equipment for video analytics modules, face recognition systems have long remained a costly pleasure. <\/p>\n

Additionally, the significant network traffic generated during operation influences the cost of the solution \u2013 beyond expenses for powerful servers, companies had to invest in active networking equipment and 'thick' communication channels.<\/p>\n

Currently, several major players in the Russian market offer high-quality algorithms for video data analysis and processing. They share an interest in projects related to large businesses. This focus is easy to explain \u2013 the cost of the solution far exceeds the capabilities of small and medium enterprises. <\/p>\n