Aligning Characteristic Descriptors with Images for Human-Expert-like Explainability

Open in new window