Interview with Laia Espuny in Cuadernos de Seguridad
Lanaccess > Articles > Interview with Laia Espuny in Cuadernos de Seguridad
  • Posted by: Demelsa Gonzalez

Cuadernos de Seguridad, one of the dean publications in the sector, in which all issues and topics related to the Security market are addressed, analyzed and studied, interviewed Laia Espuny, our Product Manager , on October 13 , who will participate at the Security Forum, next October 20, at the CCIB in Barcelona. His presentation is entitled “Intelligent Video-analytics with Deep Learning” and will take place at 1:00 p.m.

To get into the subject before the big day, he has given us some of the keys to analytical video and the use of new technologies such as Deep Learning applied to the field of security.

What are the great advantages of deep learning-based video analytics over traditional video analytics?

Broadly speaking, first-generation video analytics is limited to observing drastic changes in pixel groupings in the image. When a group of pixels with significant differences from the reference frame is detected, an intrusion alert is generated. This results in a considerable number of false alarms that operators must review individually, since any environmental, animal or vegetation factor could trigger such an alarm.

Fortunately, thanks to constant advances in graphics processing hardware  performance  and increased data availability, video analytics has evolved hand in hand with  deep learning .

The  Deep Learning  application allows new generations of video analytics to detect and identify objects and events more competently based on greater exposure to information, showing a drastic reduction in the false alarm rate compared to traditional video analytics . At the same time, these new solutions require no manual adjustments by the user and are essentially  plug-and-play , making mass deployment a realistic goal.

What are the big challenges facing video surveillance in the coming years?

The security industry constantly faces challenges. Some of them are unpredictable, such as the arrival of Covid, which has made many security systems rethink. Other challenges, on the other hand, are more predictable, even classic ones, such as bandwidth limitations. With the constant evolution of IP cameras, reaching resolutions of up to 8K, and the increase in the offer of video-analytics, the amount of data to be transmitted per second is increasing, but the bandwidth of the facilities does not grow at the same speed. That is why one of the most remarkable characteristics of Lanaccess video recorders is the efficiency and optimization in the transmission of video and other data, always evolving towards transmitting more and more, using less bandwidth.

Another challenge that will always accompany us in the sector is the need to anticipate the criminal strategy. A robbery in a bicycle shop near Madrid was recently reported in the news, causing a stir: the thieves spent 9 hours stealing material from the establishment without the intrusion alarm going off. According to the newspapers, the thieves entered dressed in white overalls and crawling, to merge with the floor of the store and deceive the presence detection system.

With a vision of the future, what will be the fields of application of video analytics? 

The main field of application, and for which the first intelligent video analytics systems were developed, is  security.  In this context, intelligent video analytics allows automatic real-time detection of risk situations, but more and more departments are adding to the advantages of adding intelligent analytics to the video surveillance system .

Beyond security, the detection, classification and object tracking capabilities offered by intelligent video analytics are allowing it to expand its fields of exploitation, acting as the  main data source for many Business Intelligence applications   used by business departments.

The set of data provided by the analytics can vary from the number of visitors in a store to the busiest areas or aisles, the average queue or waiting time, the bands with less activity, even the rate of customer recurrence. The consequent analysis of this data allows  making forecasts and isolating behavior patterns,  helping in decision-making to reduce costs and increase productivity of a business.

Another field of use that will gain weight is the  detection of the level of fraud . More and more payment facilities are digitizing ticket and ticket sales processes, such as amusement parks, museums and transportation, making it easier to buy tickets at a lower cost without meeting the necessary requirements (such as reduced tickets for minors or  seniors ). Video-analytics makes it possible to detect people who access without a ticket, as well as to estimate what percentage of visitors fits into each age group for later comparison with data on tickets sold.

November 2022
Mon
Tue
Wed
Thu
Fri
Sat
Sun
31
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
1
2
3
4






    Lanaccess necesita la información de contacto que nos proporcionas para ponernos en contacto contigo acerca de nuestros productos y servicios. Puedes darte de baja de estas comunicaciones en cualquier momento. Para obtener información sobre cómo darte de baja, así como nuestras prácticas de privacidad y el compromiso de proteger tu privacidad, consulta nuestra Política de privacidad.

    SUBE TU VÍDEO AQUÍ