iCatch Online Privacy Statement Highlights


Scope

The iCatch Online Privacy Statement applies to iCatch’s Web sites and Productions that link to the Statement. Below are some of the highlights of the Statement.

Personal Information

We collect your personal information for various purposes, such as processing your order, corresponding with you, providing you with a subscription or in connection with a job application. In addition, we may also collect information relating to your use of our Web sites through the use of various technologies, including cookies.

Uses

To fulfill your requests by us or by others involved in fulfillment. To support products or services you have obtained from us.

Your Choices

When we collect information from you, you may tell us that you do not want it used for further marketing contact. You may also turn off cookies in your browser.

News

iCatch’s SoCs are powering a diversity of camera applications in
multiple markets.

/ News /

Blog

Blog
Nov 06, 2020

AI ePTZ Webcam by V37M

 
 
Hi, this is Alex from iCatch Technology.
I am responsible for the sales and marketing of iCatch’s image processing SoC and ISP solutions. Today I am going to introduce the ePTZ webcam application powered by our AI image processing SoC - V37M. V37M is our latest SoC platform that includes ISP, video CODEC, NPU and USB 3.0 support. The ISP and video CODEC can support up to 4Kp30 for high resolution electronic pan-tilt-zoom. The NPU enables deep-learning based human detection to track the locations of the people in the camera FoV. USB 3.0 support through GPIF interface enables high resolution video transmission to the PC. The 4Kp30 capability, deep-learning based human tracking and high speed I/F makes V37M a perfect solution for ePTZ webcam application. 


Introduction

ePTZ webcam has several features to differentiate from traditional webcams. Auto-framing and presenter tracking are two good examples that we are going to talk about today. When webcam is used in a group meeting,  auto-framing will crop the image to only show the area that contains the meeting attendees. 

Take the following demo video for example:
As the meeting attendees enter the room one by one, human tracking will detect the number of attendees and their locations respectively. Then the location information is passed to the ISP to best-fit all attendees into the frame. his features makes the webcam more efficient and usable by only showing the region of interest. The presenter tracking feature utilizes the information from both human tracking and audio beamforming computed by external Audio CODEC. Audio beamforming provides the approximate location of the presenter and human tracking can determine the exact location of the presenter based on the previous information provided by audio beamforming. 

Take the following demo video for example:
With audio beamforming and human tracking, V37M can provide a PIP overlaid window with a zoomed-in view on the current presenter. This feature can help the meeting attendees to know and focus on 
the current presenter.


Live Demo

Now we have prepared a live demo to demonstrate the human tracking features explained in the previous section. The demo kit we are using today is the standard V37M SBC (Single Board Camera). SBC stands for Single Board Camera which is the standard development kit sold by iCatch for customers’ initial evaluation and prototyping. 

We have implemented deep-learning based human tracking feature in the V37M SBC. As you can see, as I move across the camera FoV, the NPU inside V37M keep tracks of my position which is highlighted by the green bonding box. This function is the fundamental technology that enables the autoframing feature. That concludes today’s introduction of V37M applied to ePTZ webcam application. 

For more information about iCatch’s products, please visit: www.icatchtek.com
For sales contact, please leave your message at: https://www.icatchtek.com/Requests