Hand Jitter Reduction Algorithm Software Test Automation Using Robotic Arm

Hand jitter reduction (HJR) is an algorithm developed to offset the jitter effect caused by camera users. The hand jitter algorithm can be added and implemented in digital signal processing using hardware acceleration to execute these noise reduction algorithms. HJR involves the detection of noise in the camera images and the application of noise reduction algorithms to smoothen these noises. HJR tests are critical for validating the authenticity of the HJR algorithm. The challenge here is to be able to simulate similar motions to what a real-world user would experience while using the camera. We have inducted and programmed a robotic arm to simulate real-world user motions. We have developed software to simulate different motions, such as the generation of sine waves, and we also generate a randomized motion that is a combination of different motions, such as a sine wave, a cosine wave, and vertical, horizontal, and angular motions. In the robotic arm setup, the robot comprises six joints, and we can rotate the joints to generate a specific motion. We have developed an algorithm to find the locations ofjoints depending on the motion we need to simulate. For example, if we need a 30° rotation while the camera is at a specific location, we can calculate the joint value for specific joints in the robotic arm. The goal of HJR tests is also to categorize the results of the camera as having acceptable or non-acceptable results based on the induced motions. The test automation has immensely helped us to objectively benchmark the performance of the algorithm over several software builds. We have provided test results of computer vision use cases of camera panorama to show the effect of hand jitter on the quality of the software.