naxtlab

*Accuracy: Probability of correctly recognizing macro block when it occurs and correctly recognizing normal screen compared to the total amount of data
*Precision: Probability of correctly recognizing macro block when it actually occurs
*Recall: Reducing probability of not recognizing macro block

The 2^nd generation engine had a low precision rate, so it occasionally recognized normal video footage as broken and sent false alarms.

Although it increased work efficiency of the IPTV service QA team by automating video quality monitoring, there was certainly room for improvement.

Our latest 3^rd generation engine has displayed a drastic improvement in the precision of macro block detection and has reduced the number of false alarms. 

Furthermore, it also increased the recall rate, which means the number of missing macro block occurrences has decreased. 

[Pictures of 3^rd Generation Engine Detecting Macro Block]

[Pictures of Comparison between 2^nd and 3^rd generation engines]
< Left: 3^rd Generation / Right: 2^nd Generation>

As you can see from the pictures above, the 3^rd generation engine successfully detected macro block even in scenes featuring grass, water, and wrinkles of clothes.

It was difficult to detect macro block correctly in these backgrounds with the 2^nd generation engine before.

Seeing those pictures really lets me feel the big step forward in technology improvement! :D

There are multiple usages of Macro Block detection technology in the Pay TV service field. 

-      Perform Aging Tests to devices

-      Broadcast service quality monitoring by Pay TV/IPTV service QA teams

-     Verify the impact of maintenance work such as improving Head End of local network

Also, you can automate your monitoring in various scenarios by using STB Multimeter solutions.

-      Round Robin monitoring

-      Freeze, Loudness simultaneous monitoring

-     Customize your alarm occurrence criteria and make it compatible with legacy system

With this Macro Block detection technology, pay TV service providers can save costs on assembling and operating specialized monitoring teams.

 If they already have monitoring teams, the macro block detection engine can also increase monitoring engineers’ work efficiency and reduce their fatigue,
as they do not have to stare at the screen for lengthy periods of time. 

Moreover, by setting alert rules, developers can focus on detecting situations where subscribers’ complaints mainly occur. 

In addition, key criteria, such as freeze, black screen, loudness, and mute, can also be monitored simultaneously.

For more details, please contact [email protected].