Patch-DrosoNet: Classifying Image Partitions With Fly-Inspired Models For Lightweight Visual Place Recognition
This addresses the need for lightweight visual place recognition algorithms for autonomous systems with budget or size constraints, representing an incremental improvement over existing DrosoNet methods.
The paper tackles the problem of visual place recognition for resource-constrained platforms by proposing a novel training approach for DrosoNet that trains separate models on distinct image regions and introduces a convolutional-like prediction method. The result is significantly improved VPR performance while maintaining an extremely compact and lightweight algorithm.
Visual place recognition (VPR) enables autonomous systems to localize themselves within an environment using image information. While Convolution Neural Networks (CNNs) currently dominate state-of-the-art VPR performance, their high computational requirements make them unsuitable for platforms with budget or size constraints. This has spurred the development of lightweight algorithms, such as DrosoNet, which employs a voting system based on multiple bio-inspired units. In this paper, we present a novel training approach for DrosoNet, wherein separate models are trained on distinct regions of a reference image, allowing them to specialize in the visual features of that specific section. Additionally, we introduce a convolutional-like prediction method, in which each DrosoNet unit generates a set of place predictions for each portion of the query image. These predictions are then combined using the previously introduced voting system. Our approach significantly improves upon the VPR performance of previous work while maintaining an extremely compact and lightweight algorithm, making it suitable for resource-constrained platforms.