Government．Smart City
Solutions
Government and Smart City
The most important issue of a Smart City is public safety. Not only because life and happiness are the groundwork of the much affirmed concept of human rights in modern history, but also because safety will attract investment, business, and skilled labors required for the development and economic growth of a city. However, as the city ushers in economic growth, it also becomes the home of crimes.
Fiction Becomes Reality
The look of a future city has been pictured for so many times in some of the best literary and cinematic works of the 20th century. Whether an Utopia or a Dystopia, they all let us speculate which technological capabilities the future might hold. By the revolutions of technology and network of the last few decades, we are now witnessing the fiction becoming reality.
Today cities are covering themselves with all kinds of “sensors” and “detectors”. And the concept for a city to use information and communication technologies (ICT) to manage its infrastructure is getting popular among urban developers.
AI Deep Learning
GVD AI Engine
Since its emergence in the 1960s, Deep Learning, as a part of Machine Learning, had been widely used in a variety of advanced systems, such as automated speech translation, automated driving, and medical research, and has produced excellent results. Then, since 2010, hardware advances have also led to more efficient methods for training deep neural networks and therefore launched the Deep Learning boom as we see today.
Unlike Machine Learning, Deep Learning has the advantages of using unstructured data to get effective training and excellent performance, which means no need for feature engineering, data labeling, and unnecessary costs.
GVD's Deep Learning uses NVIDIA's high-performance GPU to greatly speed up video image processing and significantly shorten the time of running algorithms.
PASIA™
PASIA™ means "Per-Scenarios-Self-Improved AI". It is an extended AI training designed for the Deep Learning in much complex environment or for much complex application.
PASIA™ is a highly customized service, in that, some time after sale, GVD's AI engineers will help customers to annotate the video images collected by the GVD AI server sold to form a useful dataset for Deep Learning. Then, a highly reliable, accurate, and fast neural network isn't hard to come by.
iGlance Dashboard
A Comprehensive view of Everything
iGlance Dashboard on VMS
For enterprises running business across a wide landscape, it becomes a pain in the neck to manage a large group of cameras scattered around. Now, a graphic-based "Dashboard" that visualizes every system statuses with graphs is here to ease the pain. Pies and bars let you draw conclusions quickly and provides clues that words can't. When dealing with massive data, graphics particularly show trends, gaps, and clusters.
For large projects like enterprises, government, or smart cities, iGlance Dashboard can substantially reduce the costs and labours for system maintenance and data analytics, while at the same time improves the institutions' services. For example, for a metropolitan police department, video system combining Big Data will help pinpointing crime hotspots based on where, what, and when a case was reported, and thus helps city police determine where to step up the search, inspection, or roadside check.
By using a new architecture, GVD brings an "iGlance" Dashboard to the VMS to quantify the video system's health, recording status, and number of recording days with statistics and various charts.
System iGlance
The System iGlance quickly sketches overall system statuses, including NVR Status, channel video connection status, and video recording status. Also, it auto runs graphs, such as the Channel Status Graphs, to form the images of overall channel status and lists, such as the Comprehensive Channel Status List, that breaks down every channel name and ID, recording status, video motion status, and the time last seen video-loss.
Every 60 seconds, the System iGlance auto updates these statistics: Numbers of the currently on-line but non-responsive NVRs, recording and not recoridng channels, as well as the numbers of "video-OK" and video-loss channels. A circle graph will also show to visualize how big the number of abnormal NVRs is in the system. When the mouse pointer is moved over the graph, more details become viewable.
Key data in the past seven days are summarized in bar charts, including daily alarm numbers and hourly alarm numbers. Different colors are for bars to represent different types of alarms, including: video-loss alarms, video motion alarms, CPU overheat alarms, and HDD repository damage alarms.
Application iGlance
The Application iGlance is another type of iGlance that GVD VMS features. Unlike System iGlance, which reads data from system devices, the Application iGlance interprets the data from an sophisticated application, such as License-Plate Recognition, Facial Recognition, or POS on a business floor.
Application iGlance helps you know your customers better. By analyzing shopper dwell time, dwelt shelves, and shopped items, the Application iGlance helps you discover what customers like and dislike and predict what they are going to buy before walking out. So the next time they walk in, you only have to present what they are going to buy to them. With a little adjustment to your services or shelves, maybe you can sell more, and with some lucks, maybe workforce can be reduced.
