IP Cameras: ROI Encoding Technology

HD IPCs dramatically improve the image quality, but they also bring about some bandwidth and storage issues. To resolve these issues and promote the application of HD video surveillance, Hisilicon has launched the Region of Interest (ROI) encoding solution embedded in its IPCs. This solution is developed based on the fact that a user often focuses on a particular area in the image when watching video, so the particular area in the image has to be very clear, whereas other parts can be less clear. The ROI encoding technology ensures effective surveillance over an area even when the bandwidth is insufficient.

On the monitoring image, users can select one or more ROI areas and specify unique image quality values for these areas respectively. The near-lossless compression (high code stream) technique is used for the ROI area, while the loss compression (low code stream) is used for the background or non-ROI area. This approach provides higher image quality and higher compression ratio, and resolves the problems of insufficient bandwidth and storage space.

What is self-adaptive bandwidth technology?

The network condition for video transmission fluctuates constantly within a certain range. When other network application loads are low, bandwidth available for video stream transmission increases, and vice versa. Considering the phenomenon, a camera using a fixed bit rate cannot well adapt to the changes of the network bandwidth. The bit stream may exceed the maximum bandwidth and cause network congestion, or images are not encoded in a high-quality way to use the available bandwidth maximally and wastes the network resources.

By using the self-adaptive bandwidth technology, Hisilicon network cameras can automatically adjust the video bit rate to flexibly adapt to the changes of network bandwidth. This technology helps to achieve the best image quality and the most efficient video transmission.

What is bit rate stream smoothing technology?

In a surveillance scenario, when objects in photographic field change greatly or image scenes change dramatically due to PTZ rotation, data packets of key frames (I-frames) encoded in the H.264 format increase largely in size and exceed the maximum load of the network bandwidth, which may cause network congestion and packet loss. Depending on the bit rate stream smoothing technology, Hisilicon network cameras can send peak video streams evenly during frame interval according to user configuration, avoiding packet loss caused by instantaneous data increase and ensuring the stability and reliability of video stream transmission.

What is ROI (Region of Interest) Video Surveillance Technology?

Network cameras become the mainstream in global video surveillance industry. With purpose to offer flexible and cost effective IP video surveillance systems that deliver full high definition 1080p or even 4K ultra high definition video, some manufacturers utilize many IVE technologies, and ROI is one of them. Do you know the function of ROI? ASMAG magazine interviewed Ms Cheng Wei, the marketing manager from Hikvision.

IVE stands for intelligent video encoding, IVE technologies can smartly process video data according to users demand, it can achieve optimum video surveillance performance that balanced image quality and network bandwidth, as well as storage capacity. ROI is the abbreviation of Region of Internet. It's a very useful feature in real video surveillance field. As we know, there are some regions are useless for video monitoring in captured image. These useless regions, however, occupied network bandwidth and storage. The ROI function of network camera, only concentrates on delivering high quality video from interested region. For unchanged region, it transmit less, or even doesn't transmit video.

Q: What's ROI technology, and its advantage?

A: ROI, region of interest. In real video surveillance, users typically concentrates on some specific monitored subjects, such as vehicle plate license, facial etc, however, the blue sky, lawn in the monitored area are not important. For these important regions, we called "regions of interest".

ROI encoding technology firstly established by ISO in 2000, it's one of highlight of the new generation static image compression encoding standard (JPEG2000), also it's a popular researching subject in nowadays video encoding technology. This technology can encode interested regions in video with high sufficient compression ratio. Therefore, under the same bitrate or network bandwidth, it can ensure important information will not loss, it can effectively compress image data, greatly solved the conflict between compression ratio and image quality.

Q: What's the difficulty in ROI technology?

A: Currently, there are two different ROI technologies: one is fixed region, another is dynamic tracking.

Fixed Region

In camera ROI setting page, you can turn on ROI function for fixed regions, you can select up to 4 different regions on preview image, ROI encoding level can be 1-6, adjustable.

Dynamic Tracking

Firstly, in camera facial recognition setting page, turn on facial recognition and dynamic analysis function, then go to ROI setting page to turn on dynamic tracking ROI function, ROI encoding level also can be 1-6, can achieve facial recognition and dynamic tracking ROI encoding function.

1. ROI supports triple streams technology to allow users set separately 4 fixed regions or dynamic tracking.

2. The current dynamic tracking, only supports facial recognition and ROI encoding.

ROI video surveillance triple streams

Q: In terms of cost, ROI is really worthwhile?

A: Due to the limitation of network bandwidth resources, only can use low bitrate to transmit video stream, still want to monitor important subjects clearly, ROI encoding technology can sacrifice the quality of uninterested / unimportant things, concentrate on important regions encoding, so that obtain video surveillance quality, while no need to increase bandwidth. Which in turn, significantly avoid the increasing cost of video storage and network bandwidth. Moreover, ROI technology become the common supported technology for the majority network cameras from across different manufacturers.

Q: What's the impact to video surveillance industry?

A: ROI isn't a newly developed technology. In the last year, there are major manufacturers,such as Hikvision has launched this technology, you can use this technology to solve the bandwidth shortage in some scenarios. Since the technology moves to H.265, the combination of ROI and H.265 can forwardly reduce the bitrate. By then, people will enjoy the ultra high definition video quality with low bandwidth and storage cost.