When using a deep learning network to process a large image, I break it down into smaller chunks for processing. However, when I reassemble the processed chunks, I notice significant differences in color and texture,between the adjacent areas of the chunks, making it challenging to create a coherent final image. How can I address this issue?
I attempted to introduce overlapping sections in the chunks, but the issue still persists