TK-Planes: Tiered K-Planes with High Dimensional Feature Vectors for Dynamic UAV-based Scenes
This addresses UAV-based scene perception for dynamic objects, but appears incremental as it builds on existing K-Planes NeRF methods.
The paper tackles the domain gap between synthetic and real-world data for UAV-based perception in dynamic scenes, proposing an extension of K-Planes NeRF with tiered feature vectors, and reports considerable accuracy improvements on datasets like Okutama Action and UG2 over state-of-the-art neural rendering methods.
In this paper, we present a new approach to bridge the domain gap between synthetic and real-world data for unmanned aerial vehicle (UAV)-based perception. Our formulation is designed for dynamic scenes, consisting of small moving objects or human actions. We propose an extension of K-Planes Neural Radiance Field (NeRF), wherein our algorithm stores a set of tiered feature vectors. The tiered feature vectors are generated to effectively model conceptual information about a scene as well as an image decoder that transforms output feature maps into RGB images. Our technique leverages the information amongst both static and dynamic objects within a scene and is able to capture salient scene attributes of high altitude videos. We evaluate its performance on challenging datasets, including Okutama Action and UG2, and observe considerable improvement in accuracy over state of the art neural rendering methods.