Recently, the facial recognition technology team from Dahua Technology presented their results for Labeled Faces in the Wild (LFW), which is a worldwide authoritative facial recognition database test set. With a series of technical improvements, the Dahua FaceImage facial recognition system was not only ahead of Google, Facebook, Baidu, and Tencent, it also set a new LFW record and placed Dahua's facial recognition technology in a world leading position.

Improving facial recognition

LFW was established in 2007 by the University of Massachusetts and is used to evaluate the performance of facial recognition algorithms under unconstrained conditions. It is the most widely used evaluation set in the field of facial recognition. Up until now, a few dozen teams from around the globe have provided more than 80 test results. These teams come from companies including Google, Facebook, Microsoft Research Asia, Baidu, Tencent, SenseTime, Face++, Chinese University of Hong Kong, and other leading teams from industry and academia. Many of the top-ranking methods have played a crucial role in promoting the development of facial recognition technology.

In recent years, most of the
improvement in facial recognition
technology was achieved by the
development of deep learning

In recent years, most of the improvement in facial recognition technology was achieved by the development of deep learning. Deep learning is a machine learning approach that simulates the human neural system. The functions of Convolutional Neural Networks (a commonly used deep learning model) are closely related to network depth. However, since deep networks are difficult to optimise, the network depth for facial recognition methods of the past generally ranges from a few to a few dozen layers.

High accuracy

Dahua's facial recognition technology team has designed a network with a depth of over a hundred layers (this is currently the deepest network layer amongst facial recognition system that has been announced). This enables a new type of metric learning method which allows the similarity score to be higher for images of the same person. At the same time, it lowers the similarity score between images of two different people. Combined with a highly efficient online sampling technique, the rate of convergence can also be greatly increased. By training multiple models and using a non-linear multi-model integration technique, Dahua has achieved an accuracy of 99.78% for the LFW dataset.

Dahua's facial recognition technology team is part of the advanced technology research institute in their R&D centre

Dahua's facial recognition technology team is part of the advanced technology research institute in their R&D centre, where its abundant resources in deep learning technology, hardware, data and other aspects, supported the team to deliver high performance outcome. The team supports Dahua's facial recognition related products (face detection, face feature point location, face recognition, face attribute analysis, smile detection and so on) and provides the technology to ensure robust commercial implementation.

Implementation in real-life situations

Spokesperson for the Dahua facial recognition technology team, Professor Wang Haiyang indicated that achieving high accuracy in open data sets had motivated his team to work on the more challenging implementation in real-life situations. Dahua has been a leader in the security and video surveillance industry for many years, and accumulated a huge volume of video data. By utilising this data to adjust the learning model, the performance of the algorithm in real-life situations has been greatly improved.

Furthermore, the security application requires the recognition algorithms to respond quickly. The technical team used network pruning and a multi-model feature sharing technique to greatly reduce redundant operations so that the amount of calculations required for networks with over a hundred layers is similar to that for networks with only a few dozen layers.

At the latest G20 summit, Dahua's facial recognition technology was successfully applied to enhance the security level

Application areas

Dahua's facial recognition technology has already been applied to public security, finance, and other areas. At the latest G20 summit, which has been at the centre of attention, Dahua's facial recognition technology was successfully applied to enhance the security level. Dahua cameras were deployed in core G20 activity areas and in many traffic hubs. Real-time images captured were automatically matched with security blacklists in the back-end. When a high-risk person showed up in the video, the system will issue a warning message to command centre to dispatch response immediately.

This system played an important role during the G20 summit and assisted police in arresting a number of fugitives while the system was still in its commissioning phase. Customer oriented and open innovation have been the core value of Dahua Technology. Dahua's facial recognition technology reflects its commitment to provide customers with better technology and services to implement more sophisticated solutions at lower costs.

Share with LinkedIn Share with Twitter Share with Facebook Share with Facebook
Download PDF version

In case you missed it

Security Essen 2020 cancelled due to COVID-19 pandemic
Security Essen 2020 cancelled due to COVID-19 pandemic

The international trade fair Security Essen planned for 22-25 September 2020 is cancelled due to the unforeseeable developments of the Corona crisis. Against the background of the unchanged dynamic situation and the worldwide travel restrictions, there is currently no planning security given that is required for the serious preparation of all fair participants. The customer expectations and the performance promise of the international trade fair of the security industry cannot be fulfilled under the given circumstances. The decision to cancel the fair was taken by Messe Essen in close consultation with the partner associations. Current dynamic situation The next Security Essen will take place at Messe Essen from September 20 to 23, 2022. As an international platform of the security industry, Security Essen is the trendsetting showcase for innovations every two years. Thus, numerous world pioneers had already confirmed their participation in the trade fair as well as the reputable congress and supporting programme. Our exhibitors and visitors and also we as the organisers lack the planning perspective" "The anticipation of this year's edition and the new concept of Security Essen was great among all those involved, the more difficult it was for us to decide to cancel the fair", explains Oliver P. Kuhrt, Managing Director of Messe Essen. "However, the current dynamic situation and the continuing global travel restrictions unfortunately do not permit serious preparations for the fair. As a result, our exhibitors and visitors and also we as the organisers lack the planning perspective which would have been a prerequisite for the customary professional staging of Security Essen this September.” Unclear worldwide development "The cancellation of Security Essen 2020 is the result of a prudent and careful evaluation of the current situation which we have undertaken together", adds Norbert Schaaf, Chairman of the Management Board of the Federal Association of Safety Engineering (BHE) and Chairman of the Security Essen Advisory Board. "Since the majority of our fair participants come to Essen from abroad, we had to act with particular prudence. Due to the unclear worldwide development of the Coronavirus pandemic and the associated uncertain travel planning, there was unfortunately no alternative in the end.” International fair of the security industry The cancellation of this year's edition is hitting the industry and the security associations hard" "Security Essen as an important international fair of the security industry is an absolute highlight for our association every second year", says Dr. Christian Endreß, Managing Director of the Federal Association for Security in Industry and Commerce (ASW) and member of the Security Essen Advisory Board. “The cancellation of this year's edition is hitting the industry and the security associations hard. We regret this decision, but we also fully understand it and look forward to the security year 2022 with confidence.” Enormous economic challenge Dr. Harald Olschok, General Manager of the Federal Association of the Security Industry (BDSW) and member of the Security Essen Advisory Board comments: "From the perspective of our member companies, the decision to cancel Security Essen is fully understandable. Personally, I regret this very much, because it would have been my 15th and last Security Essen as Chief Executive of BDSW and BDGW.” “But for the security service providers, the Corona crisis represents an enormous economic challenge. I am sure that the management and the entire team of BDSW and BDGW with their affiliated member companies are looking forward to Security Essen 2022 – then under better economic conditions.” The next Security Essen will take place at Messe Essen from September 20 to 23, 2022. Messe Essen and its cooperation partners are already working on the next edition and the further development of the new, innovative fair concept.

The increased role of video surveillance technology in our changing environment
The increased role of video surveillance technology in our changing environment

Today’s environment has evolved into something that according to some may seem unexplainable. But in the context of video surveillance, this is something that we understand. Allow me to shed some light and understanding in terms of security and why it truly is a necessity. Security is not a luxury, it is a necessity. An essential practice now peaking the interests of all businesses small and large. A video surveillance system is a cost effective option that does not require monitoring fees. As business slows, temporarily shuts down or closes, an increase in vacant properties is inevitable. This pandemic will continue to put severe pressure on many businesses around the country. With so many considered non-essential, it is really sad to see how many must shutter their doors and lay off employees. Keeping an eye out for suspicious activity using a commercial grade surveillance system that supports advanced analytics, may end up saving your potential customers thousands of dollars down the road. Demand for video surveillance and security products We can certainly draw on the conclusion that security is a “need” more so than a “want”. Times like this just further cement that thought process. In today’s economic spiral, people aren’t actively looking for lighting controls or home theaters. What they look for is a way to keep their loved ones safe, protect their homes, businesses and property. Video surveillance technology provides added security for you, your family and your business In my opinion, you will see video surveillance and security product sales skyrocket in the coming months and years. It has been reported that response times for first responders may be impacted as a result of COVID-19, leaving those with bad intent more time to ransack a property knowing that law enforcement may be slow to respond. Criminals will always take advantage of the situation. All we can do as a community is use common sense, stay vigilant through these odd times and watch out for one another. For some of us that may mean mitigating risks with technology. Affordable video monitoring solutions Having a solution that can quickly and securely share video footage may be the difference between identifying a perpetrator and becoming a victim. Ella, a video search platform developed by IC Realtime, makes every second of video instantly searchable and shareable, either with the authorities or your neighborhood social apps. Plus it is compatible with any RTSP streaming device. To wrap this up, it’s not about pointing out the obvious, it’s really about bringing awareness as to how technologies can be implemented to provide peace of mind without breaking the bank. Video surveillance technology is a way to do that and provide added security for you, your family and your business.

AI-based security software is the answer to vandalism and theft prevention
AI-based security software is the answer to vandalism and theft prevention

On November 2019 in Stockton, California, surveillance footage found that vandals shot out glass windows and doors in many places in a small business complex (FOX40). The intruders broke in only to leave with nothing, proving their intent was solely to vandalize the property. Meanwhile, it was reported that a trio of ATM thieves struck around 9 times across many different locations inside Brooklyn and Queens within just over a month in fall 2019 (ATM Marketplace). On average, the cost of vandalism to SMB is around $3,370 per incident (US Small Business Administration), including a staggering 692 vehicle vandalism claims per day. Likewise, the average cost of theft to SMB is about $300 per shoplifting incident and $1,500 per employee theft incident, which accounts for 38% and 34.5% of all theft instances, respectively (National Retail Security Survey). High-performance artificial intelligent systems can automate the monitoring tasks Vandalism and theft have proven time and time again to be inconvenient and deconstructively harmful towards SMB. However, these financial burdens can be prevented with the use of the right security system. AI-based security systems with Deep Learning contain many features that many SMB owners find advantageous in their pursuit to stop unwarranted and unwanted money loss.   Intrusion and loitering detection The first of many features that can help with vandalism and theft prevention is Intrusion Detection. High-performance artificial intelligent systems can automate the monitoring tasks for high-risk sites to provide a high level of security and security personnel monitoring efficiency. Traditional intrusion detection systems detect objects based on size and location, but they do not recognise the type of objects. Now, Intrusion Detection (Perimeter Protection) systems with cutting-edge, built-in AI algorithms to recognise a plethora of different object types, can distinguish objects of interest, thus significantly decreases the false-positive intrusion rate. The more advanced AI-based systems, like those we offered at IronYun, enable the users to draw ROIs based on break-in points, areas of high-valuables, and any other preference to where alerts may be beneficial. Similarly, AI Loitering Detection can be used to receive alerts on suspicious activity outside any given store. The loitering time and region of interest are customisable in particular systems, which allows for a range of detection options. Advanced loitering detection software as such can detect and trigger real-time alerts for both people loitering and/or vehicles that are illegally parked in certain areas of interest. A benefit, which only certain advanced systems contain, is the ability to send trigger actions to 3rd-party systems in reaction to receiving an alert of loitering and/or intrusion detection. These trigger actions can be set to contact authorities immediately and/or trigger a scare tactic alarm or announcement to intruder/loiterer. Certain Face Recognition and License Plate Recognition software can record individual people/vehicles Face and license plate recognition In addition to the activity detection solutions, certain Face Recognition and License Plate Recognition software can record individual people/vehicles and use pre-configured lists to identify particular faces or plates that may be of interest, such as those in watchlists. These systems can also enable the users to upload images of faces not in the lists and search for them in the camera recording. For instance, if a person is detected several times loitering outside a store, one may save one of the detection photos into the watchlist, and set up an alert when said face is recognised again outside the building in the future. The alerts will help to deter and prevent vandalism or theft, and notify the authorities to the scene before the troublemaker completes the act. The main attributes of high-performance Face Recognition systems which maximise assistance with vandalism and theft management include: Face match rate > 90% with good camera angles and lighting. Processing multiple streams and multiple faces per image. Live face extraction and matching to databases of thousands of faces within 3 seconds.                                                                                                                State-of-the-art AI security software with Deep Learning allows the user to no longer need to install special LPR camerasIf the watchlist individual is wearing a mask or their face is not in view of the camera, their license plate may be a good indicator. If a particular car is detected several times loitering in the parking lot or street outside a store, the user can set the alerts for such car to get notified in the future. With an AI solution like this, common street cameras should be equipped with LPR capabilities. So, state-of-the-art AI security software with Deep Learning allows the user to no longer need to install special LPR cameras. High-performance alert mechanisms A high-performance AI solution, in addition to having high accuracy, should be able to: Easily integrate with 3rd-party systems Work well with all ONVIF IP cameras including infrared and thermal ones (for Intrusion detection) Analyses video streams in real time and trigger alerts within a few seconds Send alerts to multiple VMSs, connect with signalling devices such as loud speakers or flashing lights Send email notifications to security staff and police departments Send notification on mobile device using AI NVR mobile app Maintains a record of all alerts to provide evidence of intrusion and loitering instances for police and insurance agencies. To assist in theft and vandalism prevention, AI-based security systems with deep learning will do all of the tedious work for you. Their low cost and high performance also make them the most accessible security solutions in the market with large return on investment. Stopping crimes is a difficult, ongoing challenge, but with the right AI software, business vendors and police departments can do so with more ease.