System and methods for image series transformation for optimal compressibility with neural upsampling
A system and method for image series transformation for optimal compressibility with neural upsampling. It incorporates a novel AI deblocking network composed of convolutional layers for feature extraction and a channel-wise transformer with attention to capture complex inter-channel dependencies. Additionally, the system includes an angle optimizer to further enhance the compressibility of an image. The convolutional layers extract multi-dimensional features from the image, while the channel-wise transformer learns global inter-channel relationships. This hybrid approach addresses both local and global features, mitigating compression artifacts and improving image quality. The model's outputs enable effective image reconstruction, achieving advanced compression while preserving crucial information for accurate analysis.
1. A system for image series transformation for optimal compressibility with neural upsampling, comprising:
a computing device comprising at least a memory and a processor;
a plurality of programming instructions that, when operating on the processor, cause the computing device to:
collect a plurality of images;
train an angle optimizer using the plurality of images on how to optimally slice the plurality of images to maximize each image's compressibility;
slice the images along an optimal angle, as determined by the angle optimizer; and
reconstruct the slices images into a plurality of reconstructed images;
encode the plurality of reconstructed images into a plurality of compressed images;
decode the plurality of compressed images into a plurality of decompressed images; and
upsample the plurality of compressed images using a multi-channel transformer, where the output includes more information than the plurality of reconstructed images.
2. The system of claim 1 , wherein the angle optimizer is a convolutional neural network.
3. The system of claim 1 , wherein the plurality of images includes medical images such as CAT scans, and aerial images which may be processed into 3D representations.
4. A method for image series transformation for optimal compressibility with neural upsampling, comprising the steps of:
collecting a plurality of images;
training an angle optimizer using the plurality of images on how to optimally slice the plurality of images to maximize each image's compressibility;
slicing the images along an optimal angle, as determined by the angle optimizer; and
reconstructing the slices images into a plurality of reconstructed images;
encoding the plurality of reconstructed images into a plurality of compressed images;
decoding the plurality of compressed images into a plurality of decompressed images; and
upsampling the plurality of compressed images using a multi-channel transformer, where the output includes more information than the plurality of reconstructed images.
5. The method of claim 4 , wherein the angle optimizer is a convolutional neural network.
6. The method of claim 4 , wherein the plurality of images includes medical images such as CAT scans, and aerial images which may be processed into 3D representations.