A ultra-high-resolution hologram generation method based on artificial intelligence and deep learning to learn fringe patterns by area

摘要

A method for generating a hologram based on artificial intelligence and deep learning that learns a fringe pattern by region, learning a deep neural network by dividing it into regions of a hologram, properly connecting and accumulating fringe patterns obtained for each region, and generating a hologram, (a) A first fringe pattern generated by a computer generated hologram (CGH) algorithm for points of a three-dimensional object, and a second fringe pattern by a fringe pattern generator, and the first and second fringe patterns learning the fringe pattern generator by optimizing the parameters with a probability value of a similar degree, setting conditions to the position of the object point and the hologram area corresponding thereto; and, (b) generating third fringe patterns for each region through the learned fringe pattern generator, multiplying the generated third fringe pattern for each region by a brightness value, accumulating the multiplied value, and generating a third fringe pattern for each region Concatenating to prepare a configuration including the step of generating a digital hologram. In the same way as described above, by dividing the hologram into regions and learning the deep neural network, and by appropriately connecting and accumulating the fringe patterns obtained for each region, the fringe pattern can be divided and generated for easy processing even for a large-sized super-resolution hologram. can