Storage management methods and systems

A method for image processing comprising identifying image locations corresponding to a read path 112; generating, from the location data, parameter data relating to a characteristic of the read path within pixel blocks (pixel tiles) in the image (e.g. see surface areas between trajectory 112 and the lower boundaries of each of blocks 114b, 114c, 114b etc); and based on the parameter data assigning different storage prioritisation values to the various blocks (see table 1, p. 14). In this way, blocks with low prioritisation values may be evicted from the cache while blocks more likely to be needed again may be assigned a high priority for retention in the cache. The surface area between the read trajectory and the lower boundary of the block may be calculated by approximation (fig 4). This method may be more efficient than a Least Recently Used (LRU) algorithm. The storage may be a cache memory of an image processing system (GPU etc).