English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

How can we localize ourselves within a building solely using visual information, i.e., when no data about prior location or movement are available? Here, we define place categorization as a set of three distinct image classification tasks for view matching, location matching, and room matching. We present a novel image descriptor built on texture statistics and dynamic image partitioning that can be used to solve all tested place classification tasks. We benchmark the descriptor by assessing performance of regularization on our own dataset as well as the established Indoor Environment under Changing conditionS dataset, which varies lighting condition, location, and viewing angle on photos taken within an office building. We show improvement on both the datasets against a number of baseline algorithms. INTRODUCTION Humans possess the remarkable ability to reliably localize themselves in the world under a multitude of conditions: Indoors and outdoors, in unknown terrain, with reduced senses during different weather conditions, and even under cognitive load while processing other tasks. But how do we solve this problem? Answering this question will not only give crucial insight into mechanisms of spatial cognition in the human brain but may also be important for selflocalization of mobile robots. For humans, the dominant sensory information used for place recognition is vision [1]. However, the mechanisms that lead from visual input to knowledge of a place are poorly understood. What are the features we look at to determine where we are? Or to detach the question from its anthropological context: What are the features that best discriminate between places? Optimal features for this task would have to be specific for a certain place but invariant to different views within the location [2]. It is important to distinguish these requirements from the requirements for vision-based Simultaneous Localization and Mapping (SLAM; see Section 1.1 for a more detailed discussion). The features required there (tracking features) need to be stable across successive camera frames and recognizable from slightly different viewpoints, whereas the features suitable for vision-only place recognition have much stronger invariance requirements. For example, place features need to be matched across completely different views from a place, and they need to be stable over long periods of time with varying illumination conditions. Consequently, successfully matching place features is a much harder task than matching tracking features. But what are suitable place features for this task? On larger scales, simple, global image statistics have been found to be discriminative between places. In particular, a global texture histogram descriptor has been shown to work for place classification tasks ranging from city-scale to world-scale localization [3]. For indoor localization [4], also use histograms over relatively simple orientation descriptors on the Indoor Environment under Changing conditionS (INDECS [5]) database. However, these features are very generic, and confusion may arise if the same texture feature is found in multiple areas of an image, e.g., at the ceiling and on the floor. In this paper, we present an approach that circumvents this problem by partitioning the image according to texture occurrences.

Indoor place categorization based on adaptive partitioning of texture histograms