These tools and metrics are designed to help AI actors develop and use trustworthy AI systems and applications that respect human rights and are fair, transparent, explainable, robust, secure and safe.
The structural similarity index measure (SSIM) measures the perceived similarity of two images. When one image is a modified version of the other (e.g., if it is compressed) the SSIM serves as a measure of the fidelity of the compressed representation. The SSIM formula is based on three comparison measurements associated with luminance, contrast, and structure.
SSIM can indirectly support 'Safety' and 'Robustness' in AI systems that generate or process images. High structural similarity between expected and actual outputs may help ensure that critical visual information is preserved, reducing the risk of harmful or misleading outputs (Safety). Additionally, monitoring SSIM under varying conditions can help assess the system's ability to maintain performance when faced with noise or perturbations (Robustness). However, these connections are indirect and context-dependent, as SSIM alone does not guarantee safety or robustness without broader system-level considerations.
References
About the metric
You can click on the links to see the associated metrics
Objective(s):
Lifecycle stage(s):
Target users:
Github stars:
- 7100
Github forks:
- 720
