The Hidden Order Framework: From Evidence-Based Prediction in Scenario Spaces to Prompt Learning in Image Production Processes

The event scenario interpretation effectiveness in images is strongly related to the perceptual computation of image plane complexity. In order to interpret event scenario behaviors in multi-temporal manual mapping process, a reversible information hiding algorithm for predicting image neighborhood geo-features in exclusive-or (XOR) scrambling framework is proposed in this paper which can identify different objects, behaviors, and scenarios, and support image event complexity measurements. The xor scrambling decryption can simultaneously protect the statistical information and location information of the original image, make up for the semantic missing problem in the image prediction stage, and enhance the discriminative efficiency of the prediction of the hidden events in the image scenarios. In this paper, the characteristics of remote sensing images, planar landscape images and scrambled image semantics as well as the digital divide among them are reviewed and discussed, and a “scenario space in the xor scrambling framework” is proposed to cope with the divide. Then, an evidence-based ubiquitous knowledge mining approach is presented to realize the work of spatial location information superposition through “hidden order” mining experiments. Finally, we demonstrate the application feasibility of the model. In summary, this method can achieve context-oriented implicit knowledge association among remote sensing images, landscape paintings and image semantics, and put forward high-level, cross-disciplinary, cross-network and multi-modality collaborative viewing requirements, which can help to improve the transition from the end of “image interpretation + deep learning” to the end of “ prompt learning + image production”.