16-bit integer multiplication and a very small amount of 32-bit multiplication and division. In this way, our
early digital cameras were possible. Secondly, the various template processing common in PS image processing
also uses the two-dimensional matrix convolution method, which also requires a huge amount of integers (8-
bit image image size requires 16-bit and 32-bit intermediate calculation results) Multiplication Adding
calculations, using discrete convolution will greatly improve the operation speed. Therefore, the STC8 single-
chip microcomputer with MDU can not only be used to collect and display images in real time, but also process
images in real time. Finally, artificial intelligence also involves a large number of vector and matrix operations,
such as neural network convolution, which can be implemented with the combination of function 4 and function
6. MDU should be able to be applied in small intelligent scenarios. Just to realize these functions, it needs the
cooperation of the enhanced double data pointer of STC8, and needs a special knowledge structure, and a
function library is specially compiled to provide users with the huge advantage of STC8 MDU.
To Next Page
Loading...