<p>This paper analyzes the saliency indicators representing human visual attention and proposes a weighted mean square deviation local dimming method based on an image saliency ranking model. It references the saliency ranking model for instance segmentation and saliency ranking, determining the saliency levels of different salient objects in the image, which reflect the importance of different regions. Then, it assigns weights to different regions in the image based on their saliency levels, with regions having higher weights being considered more critical in the image. During dimming, details of critical regions are preserved while the image quality of low attention areas is appropriately reduced, thereby enhancing the subjective visual effects of the image and lowering the power consumption of liquid crystal displays. Specifically, this paper determines the importance of different regions in the image based on saliency order and selects appropriate methods to weight the allowable saturation error in mean square error calculation, ensuring that the optimized saturation error is within the human perceptual range, thereby determining the final backlight values for each region. Simulation results indicate that, compared to common local dimming algorithms and saliency-based global dimming algorithms, an optimal balance is achieved when considering image quality, power consumption, and general applicability across various scenarios.</p>