Addressing the Depth-of-Field Constraint: A New Paradigm for High Resolution Multi-Focus Image Fusion
This work addresses the problem of generating sharp images from multiple focus points for applications in photography and imaging, representing a new paradigm in the field.
The paper tackles the depth-of-field limitation in multi-focus image fusion by proposing VAEEDOF, a method using a distilled variational autoencoder, and introduces the MattingMFIF dataset to address data scarcity, achieving state-of-the-art results with seamless artifact-free fused images.
Multi-focus image fusion (MFIF) addresses the depth-of-field (DOF) limitations of optical lenses, where only objects within a specific range appear sharp. Although traditional and deep learning methods have advanced the field, challenges persist, including limited training data, domain gaps from synthetic datasets, and difficulties with regions lacking information. We propose VAEEDOF, a novel MFIF method that uses a distilled variational autoencoder for high-fidelity, efficient image reconstruction. Our fusion module processes up to seven images simultaneously, enabling robust fusion across diverse focus points. To address data scarcity, we introduce MattingMFIF, a new syntetic 4K dataset, simulating realistic DOF effects from real photographs. Our method achieves state-of-the-art results, generating seamless artifact-free fused images and bridging the gap between synthetic and real-world scenarios, offering a significant step forward in addressing complex MFIF challenges. The code, and weights are available here: