Aggregating Multiple Bio-Inspired Image Region Classifiers For Effective And Lightweight Visual Place Recognition

Bruno Arcanjo,Bruno Ferrarini,Maria Fasli,Michael Milford,Klaus D. McDonald-Maier,Shoaib Ehsan

2023-12-20

Abstract:Visual place recognition (VPR) enables autonomous systems to localize themselves within an environment using image information. While VPR techniques built upon a Convolutional Neural Network (CNN) backbone dominate state-of-the-art VPR performance, their high computational requirements make them unsuitable for platforms equipped with low-end hardware. Recently, a lightweight VPR system based on multiple bio-inspired classifiers, dubbed DrosoNets, has been proposed, achieving great computational efficiency at the cost of reduced absolute place retrieval performance. In this work, we propose a novel multi-DrosoNet localization system, dubbed RegionDrosoNet, with significantly improved VPR performance, while preserving a low-computational profile. Our approach relies on specializing distinct groups of DrosoNets on differently sliced partitions of the original image, increasing extrinsic model differentiation. Furthermore, we introduce a novel voting module to combine the outputs of all DrosoNets into the final place prediction which considers multiple top refence candidates from each DrosoNet. RegionDrosoNet outperforms other lightweight VPR techniques when dealing with both appearance changes and viewpoint variations. Moreover, it competes with computationally expensive methods on some benchmark datasets at a small fraction of their online inference time.

Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

The paper aims to address key issues in the field of Visual Place Recognition (VPR), particularly focusing on achieving efficient and accurate visual localization on resource-constrained hardware platforms such as mobile robots. Specifically, the goals of the paper can be summarized as follows:

1. **Improve computational efficiency**: Current methods based on Convolutional Neural Networks (CNN) perform well in VPR but have high computational demands, making them unsuitable for low-end hardware platforms. Therefore, one of the goals of the paper is to enhance the performance of VPR systems while maintaining low computational complexity.

2. **Enhance VPR system robustness**: Existing lightweight VPR technologies often fail to ensure sufficient robustness and accuracy when facing challenges such as lighting changes, seasonal changes, viewpoint changes, and dynamic elements in the scene. The paper proposes a new multi-DrosoNet localization system (RegionDrosoNet) aimed at improving VPR performance, especially in handling appearance and viewpoint changes.

3. **Design lightweight VPR methods**: The paper introduces a new model called DrosoNet, a small neural network model inspired by the olfactory processing of fruit flies. By combining multiple DrosoNet units, leveraging their inherent randomness to increase model diversity, and introducing a novel voting module to aggregate the outputs of these units, the paper aims to achieve efficient VPR tasks.

4. **Achieve model differentiation**: To further enhance the performance of VPR systems, the paper proposes a new method—RegionDrosoNet. This method increases external model diversity by dividing the input image into different regions and training specialized groups of DrosoNet units to handle these regions.

5. **Optimize decision-making mechanisms**: The paper also proposes an improved voting mechanism that considers not only the top-scoring candidate reference location of each DrosoNet unit but also multiple top reference candidates. This helps the system converge to the most widely recognized reference location, reducing the impact of potential mismatches produced by individual DrosoNet units.

In summary, the main goal of the paper is to develop an efficient and accurate VPR solution, particularly on resource-constrained hardware platforms, to meet the practical application needs of mobile robots.