To decrease the analytical complexity of the exponential feedback in the classical Chialvo map, a structurally Simplified Chialvo Map (SCM) is proposed in this work. The SCM maintains the foundational dynamics of the original Chialvo map. By replacing the exponential term with a linear term and after a normalization-based suppression, the Improved Simplified Chialvo Map (ISCM) is obtained. It is found that the ISCM preserves the original topological features while exhibiting higher local instability, as validated by entropy-based measures. Based on the chaotic properties, an encryption framework integrating Convolutional Neural Networks (CNNs) is developed with the ISCM model along with the tiny encryption algorithm. Within this scheme, CNN-extracted features generate content-adaptive masks as a nonlinear security barrier, strengthening resistance to known-plaintext attacks. By leveraging the high entropy and low computational cost of the ISCM, the proposed scheme demonstrates superior performance in terms of both encryption efficiency and security.
Yang et al. (Fri,) studied this question.