GravNav: Using a Gravity Model for Multi-Scale Navigation Waqas Javed, Sohaib Ghani, and Niklas Elmqvist Purdue University West Lafayette, USA {wjaved, sghani, elm}@purdue.edu ABSTRACT We present gravity navigation (GravNav), a family of multi-scale navigation techniques that use a gravity-inspired model for assist- ing navigation in large visual 2D spaces based on the interest and salience of visual objects in the space. GravNav is an instance of topology-aware navigation, which makes use of the structure of the visual space to aid navigation. We have performed a controlled study comparing GravNav to standard zoom and pan navigation, with and without variable-rate zoom control. Our results show a significant improvement for GravNav over standard navigation, particularly when coupled with variable-rate zoom. We also report findings on user behavior in multi-scale navigation. Keywords Topology-aware navigation, guided navigation, zooming, panning, multi-scale spaces, space-scale diagrams. Categories and Subject Descriptors H.5.2 [Information Interfaces and Presentation]: User Interfaces— Interaction styles; I.3.6 [Computer Graphics]: Methodology and Techniques—Interaction techniques 1. INTRODUCTION Human attention has many similarities to the theory of gravity: regions of high visual interest attract attention just like moths are drawn to an open flame at night, and interest wanes with distance. For example, the concept of visual salience as a measure of the per- ceptual distinctiveness of an object is key in directing our visual at- tention [17, 23, 25]. Furthermore, Fisher [8] found that map atten- tion is centered on coastlines, most likely due to their irregularity and corresponding high visual interest. Despite these facts, it is cu- rious that so few navigation techniques for computer applications take advantage of this phenomenon. Standard scrolling, panning, and zooming techniques all disregard the structure and topology of the underlying visual space, essentially making it equally easy (or equally difficult) to navigate to an empty cornfield on an interactive map as it is to navigate to an urban area. Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. AVI ’12 Capri, Italy Copyright 2012 ACM 978-1-4503-1287-5 ...$10.00. In this paper, we attempt to remedy this state of affairs by pre- senting the concept of gravity navigation (GravNav) where an interest-based gravity model is used to ease navigation in large vi- sual spaces. The basic idea is to compute an attention-gravity vec- tor given the user’s location in the visual space and the surrounding items of interest, and then use this vector to guide navigation. Sim- ilar to semantic and sticky pointing [4, 6, 20], the intuition is to speed up navigation over empty areas, and to slow it down around regions of interest. In particular, we believe the technique is espe- cially well-suited for assisting micro-level navigation in multi-scale spaces [11], where small disparities at a distant magnification level may have massive impact when zooming into the space to see de- tails and interact with visual objects. The applicability of this family of gravity navigation techniques encompass all interactive applications that incorporate large vi- sual spaces, such as interactive maps (e.g., Google or Bing Maps), large-scale zoomable visualizations, or detailed CAD blueprints and drawings, etc. To validate the utility and efficiency of grav- ity navigation, we performed a controlled user study for multi- scale navigation where we compared performance for two instan- tiations of the concept—gravity zooming and panning—with the standard zooming and panning operations that would be found to- day in the above interactive applications. We also included the Or- thoZoom [2] zoom-gain interaction technique as a condition in our experiment. Our results are convincingly in favor of the new tech- nique, showing significantly decreased completion times across all navigation scales and distances. We also report on navigation be- havior for different conditions using space-scale diagrams [11], showing that gravity navigation is particularly suited for combin- ing with OrthoZoom due to the reduced need for clutch operations. 2. RELATED WORK Our work here draws upon existing work on general and multi- scale navigation, assisted navigation, and pointing. 2.1 General and Multi-Scale Navigation Pan and zoom are the canonical navigation operations [11, 22], allowing the user to change the position and the magnification of the viewport, respectively. Scrolling is a special instantiation of panning restricted to one dimension, such as for navigating in large documents. Much work has been dedicated to exploring effective pan and zoom operations, including space-scale diagrams for mod- eling navigation behavior [11], optimal tradeoffs between zoom and pan [24], and effective methods for view navigation [10]. Most pan and zoom operations take place in multi-scale spaces, i.e., spaces that contain objects with scale-dependent appearances. For example, a map database provides more detail as the user zooms from a world view down to a city, neighborhood, or street. 1