In this study, we introduce a novel method for enhancing diffusion images through a hybrid partial differential equation approach. This approach incorporates pulse filtering, an enhanced forward-backward diffusion filter, and the total variational diffusion method. Our algorithm effectively enhances small boundaries in images by utilizing an improved hybrid partial differential equation diffusion image filtering algorithm. By utilizing a pulse filter with specific parameter constraints, we enhance the processing capabilities for strong boundaries. Additionally, we employ the level set method to eliminate sawtooth artifacts at boundaries and enhance boundary smoothness. Furthermore, we propose a biorthogonal filtering method based on a degraded model that achieves biorthogonal mapping in the spatial domain, ensuring the invertibility of the degraded extended image concerning the degraded model. Simulation results demonstrate the superiority of our method compared to other techniques. The resulting amplified images exhibit excellent photorealism, with improved boundary details for both weak and medium brightness levels. This method effectively preserves image boundary details while effectively reducing noise.