| Summary: | CNNs have excelled at performing place recognition over time, particularly when the neural network is optimized for localization in the current environmental conditions. In this paper we investigate the concept of feature map filtering, where, rather than using all the activations within a convolutional tensor, only the most useful activations are used. Since specific feature maps encode different visual features, the objective is to remove feature maps that detract from the ability to recognize a location across appearance changes. Our key innovation is to filter the feature maps in an early convolutional layer, but then continue to run the network and extract a feature vector using a later, more viewpoint invariant layer in the same network. Our approach improves the condition and viewpoint invariance of a pre-trained network, using as little as a single training image pair from the deployment environment. An exhaustive experimental analysis is performed to determine the full scope of causality between early layer filtering and late layer extraction. For validity, we use three datasets: Oxford RobotCar, Nordland, and Gardens Point, achieving overall superior performance to NetVLAD. The work provides a number of new avenues for exploring CNN optimizations, without requiring any re-training of the network weights.
|
|---|