The proliferation of counterfeit currency is a global challenge that poses a significant threat to the stability of economies worldwide. It undermines trust in paper currency which affects financial systems and transactions. In this paper, we present a novel multi-input single-output deep learning approach for reliable banknote authentication. Our method assesses every crucial security parameter across multiple segments of the banknote encompassing both front and back faces which effectively addresses the challenges associated with blurriness during image capturing. To implement our system, we have utilized transfer learning with popular pre-trained models and conducted rigorous hyperparameter optimization and fine-tuning. Our DenseNet121-based model has achieved a remarkable test accuracy of 98.9% and an average inference time of 350 ms. Our research has the potential to enhance public trust in paper currency and mitigate the impact of money inflation. By effectively combating counterfeiting, our proposed model contributes to economic stability and secures financial transactions, making it a valuable tool for various industries and applications.