Your Camera Algorithms Matters!

Nowadays, more and more intelligent functions are embedded in smart cameras. Object detection, a computer vision technology, is one of the critical camera algorithms used to detect and recognize the visually observable objects in video streaming, such as human, human faces, cars, and obstacles (e.g. trees, buildings, utility poles) etc. Once camera algorithms fail to detect or use the wrong formula, it will trigger the potential risks which may leave negative impacts on user experiences and further destroy the company’s brand reputation.

The Risks of Your Smart Cam

  • Home property safety: a security camera fails to detect a burglar breaking in
  • Personal safety in an autonomous car: a camera fails detect traffic signals accurately and lead a car accident
  • User doesn’t get alert when cloud algorithms identify moving objects or unusual activities

“Simulation Engine Testing Solution”

A Cost-Efficient and Effective Solution to Camera Software Issues

To pinpoint and fix the camera issues in different situations, you had to build up various real scenarios to test camera software. But now, with Allion’s new launched “simulation engine testing solution,” it is allowed to select any VR (Virtual Reality) like scene to verify camera algorithms effectively and efficiently. Our testing scenarios can be customized designed according to demanding applications, such as web camera for home security, dash cam for automotive. The solution can be quickly implemented in the product development process, from EVT, DVT to PVT stage.

◆ A powerful Scenario-based Library: various testing environments/ scenarios and control variables

◆ AI, VR Simulator: 

  • Create what-you-see is what-you-use environment by simulating the real world
  • Reproduce various scenarios with control variables in a short time and discover critical issues through repetitive tests
  • The accumulated issue database “Simulation Engine Library” is beneficial to further in-depth analysis

 

◆ Time saving and cost effective for camera makers and their APP developers: save time and money to build up real testing environments

◆ A Trustworthy Testing Method: the test results of real world scenario are in accordance with those of simulation engine.

Service Target

Test Environment & Equipment

Scene Settings

Pre-define the control variables to fit in the five Test Fields, such as brightness, weather (outside), humans, walking speed, route, frequency, duration …etc

Test Scope

1. Function Test & Benchmark Test

Function testing is to test sensor sensitivity and accuracy of object detection function in the different scenarios. The competitive analysis report will help vendors to know the strengths, weaknesses, and potential risks of its product by comparing other competitor’s products.

2. Durability Test

To test whether camera’s object detection could function well or not in different time intervals.

  • Short-term Test
  • Long-term Test: 24/7
3. Wireless Network Test

To test camera’s object detection could function well or not in different wireless conditions.

  • Interference Test: Bluetooth®, Microwaves, CCI, ACI
  • Wi-Fi Signal Quality Test
4. Interoperability Test

To test the camera can perfectly work with AP and mobiles.