In recent years, explainable machine learning has become increasingly important for ensuring transparency and trust in machine learning systems, particularly in high-stakes applications such as medical imaging and autonomous systems. However, existing explanation methods for image classification models often provide explanations that vary in accuracy and do not adapt to user-specific needs or constraints. This project aims to develop a novel algorithm that enables users to control the fidelity level of the generated explanations for image classification by introducing a user-adjustable parameter that modulates the granularity and completeness of visual attributions. The algorithm offers flexible trade-offs between interpretability and computational cost. The proposed method builds on state-of-the-art attribution techniques and incorporates fidelity-aware regularization to maintain explanation faithfulness. The approach empowers users, such as domain experts and developers, to tailor explanations to their context, enhancing the usability and trustworthiness of machine learning systems.