WiMi Hologram Cloud Inc. (NASDAQ: WIMI) (“WiMi” or the “Company”), a leading global Hologram Augmented Reality (“AR”) Technology provider, today announced the development of a graphics processor data fusion algorithm system.
The GPU data fusion algorithm system improves the general-purpose computing capabilities and the interoperability of the GPU. Using the algorithmic computing power of the GPU, the algorithm enhances the versatility and task provisioning capabilities. The algorithm computes and processes data from multiple sources that are spatially or temporally redundant or complementary to obtain more accurate and richer information than any single data, resulting in a composite image with new spatial, spectral, and temporal characteristics. The algorithm can use multi-faceted, multi-layered data processing sensors to get more meaningful information than that from a single sensor.
WiMi’s system uses the main filter to compare the weighting coefficients of each local filter with the set threshold. In the case of weighting coefficients below the threshold, no feedback correction is made. In the case of weighting coefficients above the threshold, feedback correction is made by the main filter. The system fuses large amounts of multi-source data from multiple sensors for graphical processing. Stable and reliable output perception results can be obtained in different environments, enabling self-adaptation of environment perception and enhancing the system’s robustness. For example, the system can be enhanced for improved classification of images, confidence, and reduced ambiguity. In extending the range of observations in time and space, the system can enhance the confidence and discrimination of data, improving reliability, ability to describe the environment, and speed of information processing. The system can be used in applications such as linear algebra, fluid simulation, data analysis, and scientific operations that require large amounts of redundant data set processes and intensive memory access.
WiMi uses a suitable algorithm to combine two or more source images to obtain a new idea. Enhancing the useful information in the image makes it possible to increase the accuracy of image understanding and acquire more accurate results. In multiple fusion applications, WiMi sets different levels of multi-source data to be processed in this system, each representing an extra level of data abstraction, depending on the purpose of the application, the input data being processed, and the degree of data pre-processing before fusion processing. Image information fusion is divided into three levels from low to high, namely pixel-level fusion, feature-level fusion, and decision-level fusion, according to the different levels of information abstraction. The algorithm used and the scope of application vary according to the fusion level.
The system provides image fusion of complementary and redundant information in each image data, which is more in line with the visual characteristics of humans and machines and helps people to analyze the image in depth and to detect, identify or track the target effectively. Based on this, WiMi continues to study the effective spatially accurate configuration of image data information acquired by each type of sensor in the same scene. Based on this, the valuable information of various image data is complemented. At the same time, this data information is effectively fused to make new, high-quality, huge image information.