Implicit Neural

Implicit neural representations (INRs) are a novel approach to representing data, such as images and videos, using the parameters of a neural network, aiming for efficient storage and processing. Current research focuses on improving encoding and decoding speeds, enhancing compression performance through techniques like parameter reuse and decomposition of static and dynamic information, and extending INRs to handle diverse data types and tasks including video compression, super-resolution, and even spatiotemporal traffic data modeling. This approach offers significant potential for reducing storage requirements and computational costs in various applications, particularly in areas dealing with high-dimensional data like video and 3D scene representation.