Point-feature Label Placement Research Articles

Reinforced Labels: Multi-Agent Deep Reinforcement Learning for Point-Feature Label Placement.

Over the recent years, Reinforcement Learning combined with Deep Learning techniques has successfully proven to solve complex problems in various domains, including robotics, self-driving cars, and finance. In this paper, we are introducing Reinforcement Learning (RL) to label placement, a complex task in data visualization that seeks optimal positioning for labels to avoid overlap and ensure legibility. Our novel point-feature label placement method utilizes Multi-Agent Deep Reinforcement Learning to learn the label placement strategy, the first machine-learning-driven labeling method, in contrast to the existing hand-crafted algorithms designed by human experts. To facilitate RL learning, we developed an environment where an agent acts as a proxy for a label, a short textual annotation that augments visualization. Our results show that the strategy trained by our method significantly outperforms the random strategy of an untrained agent and the compared methods designed by human experts in terms of completeness (i.e., the number of placed labels). The trade-off is increased computation time, making the proposed method slower than the compared methods. Nevertheless, our method is ideal for scenarios where the labeling can be computed in advance, and completeness is essential, such as cartographic maps, technical drawings, and medical atlases. Additionally, we conducted a user study to assess the perceived performance. The outcomes revealed that the participants considered the proposed method to be significantly better than the other examined methods. This indicates that the improved completeness is not just reflected in the quantitative metrics but also in the subjective evaluation by the participants.

A Hybrid Discrete Artificial Bee Colony Algorithm Based on Label Similarity for Solving Point-Feature Label Placement Problem

The artificial bee colony algorithm (ABC) is a promising metaheuristic algorithm for continuous optimization problems, but it performs poorly in solving discrete problems. To address this issue, this paper proposes a hybrid discrete artificial bee colony (HDABC) algorithm based on label similarity for the point-feature label placement (PFLP) problem. Firstly, to better adapt to PFLP, we have modified the update mechanism for employed bees and onlooker bees. Employed bees learn the label position of the better individuals, while onlooker bees perform dynamic probability searches using two neighborhood operators. Additionally, the onlooker bees’ selection method selects the most promising solutions based on label similarity, which improves the algorithm’s search capabilities. Finally, the Metropolis acceptance strategy is replaced by the original greedy acceptance strategy to avoid the premature convergence problem. Systematic experiments are conducted to verify the effectiveness of the neighborhood solution generation method, the selection operation based on label similarity, and the Metropolis acceptance strategy in this paper. In addition, experimental comparisons were made at different instances and label densities. The experimental results show that the algorithm proposed in this paper is better or more competitive with the compared algorithm.

A point-feature label placement algorithm based on spatial data mining.

The point-feature label placement (PFLP) refers to the process of positioning labels near point features on a map while adhering to specific rules and guidelines, finally obtaining clear, aesthetically pleasing, and conflict-free maps. While various approaches have been suggested for automated point feature placement on maps, few studies have fully considered the spatial distribution characteristics and label correlations of point datasets, resulting in poor label quality in the process of solving the label placement of dense and complex point datasets. In this paper, we propose a point-feature label placement algorithm based on spatial data mining that analyzes the local spatial distribution characteristics and label correlations of point features. The algorithm quantifies the interference among point features by designing a label frequent pattern framework (LFPF) and constructs an ascending label ordering method based on the pattern to reduce interference. Besides, three classical metaheuristic algorithms (simulated annealing algorithm, genetic algorithm, and ant colony algorithm) are applied to the PFLP in combination with the framework to verify the validity of this framework. Additionally, a bit-based grid spatial index is proposed to reduce cache memory and consumption time in conflict detection. The performance of the experiments is tested with 4000, 10000, and 20000 points of POI data obtained randomly under various label densities. The results of these experiments showed that: (1) the proposed method outperformed both the original algorithm and recent literature, with label quality improvements ranging from 3 to 6.7 and from 0.1 to 2.6, respectively. (2) The label efficiency was improved by 58.2% compared with the traditional grid index.

A point-feature label placement algorithm considering spatial distribution and label correlation

A point-feature label placement algorithm considering spatial distribution and label correlation

Point feature label placement for multi-page maps on small-screen devices

Map applications on mobile devices such as smartphones and smartwatches have become ubiquitous. When visualizing spatial data on such small-screen devices, one major challenge is annotating the data with labels (e.g., small icons). The restricted space requires new visualization techniques as established strategies, such as maximizing the number of placed labels, easily lead to the omission of information. We propose an approach that distributes all labels contained in a temporarily fixed map section on multiple pages. Applying interaction techniques for navigating through the pages, a user can access all information both without any overlapping labels and without the need for zooming. We propose a method with two phases; a pre-processing phase and a query phase. We use an optimization approach to pre-compute labelings on the level of a whole city and provide the on-demand querying of individual labelings at a more local level. Our approach provides a consistent label-page assignment, meaning that labels do not appear and disappear when the user pans the map. Further, our model provides quick access to important information and a spatially balanced distribution of labels on pages. In experiments on real-world data we analyze different parameter settings and show that our model yields high-quality labelings.

A parallel annotation placement method for dense point of interest labels using hexagonal grid

ABSTRACT Point feature label placement (PFLP) has been a fundamental problem in automatic cartography over decades. In this research, labels are approximated by individual characters to offer more freedom of placement on raster-based grids, including allowing labels turning to avoid occlusions. Considering that parallel orientation of label annotation increases the label chances to resolve the dense conflicts, and that hexagonal grids provide more orientations than square grids, this study uses a parallel annotation to conduct point label placement on hexagonal grids, which is efficient to figure out high-quality label placement for points of interest (POI). The approach introduces a detailed representation of point feature labels on hexagonal grids. Then an algorithm of label placement on hexagonal grids is presented, using the methodology of parallel annotation to arrange labels with different combinations of label orientations. The approach is fast and flexible due to the pre-computed grid and maintains good readability due to the parallel annotation.

Towards unambiguous map labeling - Integer programming approach and heuristic algorithm

The development of computer systems that emulate human expertise in decision making is a common task in the automation of processes like planning, design, prediction or control. When it comes to map design, one of the problems of the cartographer is to orchestrate the elements of the map so that it clearly conveys its information to the readers. Several disciplines, such as Computational Geometry, Cartography and Operations Research have devoted diverse efforts to automate this task, commonly identified in the research domain as the Point-Feature Label Placement (PFLP). These approaches aim to avoid or minimize overlapping elements on the map, but almost no attempt to model possible ambiguities of feature and label pairs has been conducted so far. This article presents four Integer Programming formulations for the PFLP problem that incorporate ambiguity as an indicator of the solution quality. A very basic heuristic procedure is also introduced as an alternative to the exact formulations for large instances. All the proposed methods are tested on several maps of Spanish municipalities. Finally, some conclusions and recommendations are given.

A Labeling Model Based on the Region of Movability for Point-Feature Label Placement

Automatic point-feature label placement (PFLP) is a fundamental task for map visualization. As the dominant solutions to the PFLP problem, fixed-position and slider models have been widely studied in previous research. However, the candidate labels generated with these models are set to certain fixed positions or a specified track line for sliding. Thus, the whole surrounding space of a point feature is not sufficiently used for labeling. Hence, this paper proposes a novel label model based on the region of movability, which comes from plane collision detection theory. The model defines a complete conflict-free search space for label placement. On the premise of no conflict with the point, line, and area features, the proposed model utilizes the surrounding zone of the point feature to generate candidate label positions. By combining with heuristic search method, the model achieves high-quality label placement. In addition, the flexibility of the proposed model enables placing arbitrarily shaped labels.

A Comprehensive Multi-criteria Model for High Cartographic Quality Point-Feature Label Placement

The lettering process, including assigning names to point features, is an essential part of map production. While there have been numerous and varied research efforts to automate point-feature label placement (PFLP), none of them seems to have taken into account the many well-established cartographic precepts for point-feature annotation used by human cartographers. As a result, current fully automated solutions are limited in their expressive power. The PFLP problem is still vital, therefore, and solving it is a compelling challenge. This article presents a comprehensive multi-criterion model that complies with almost all well-defined cartographic placement principles and requirements for PFLP, allowing for a significant increase in toponym density without affecting legibility. The proposed model, expressed as a quality-evaluation function, can be used by any mathematical optimization algorithm to resolve the automated label-placement problem. Through an application of the proposed model tested on volunteered geographic (VGI) data and the creation of sample parameter settings, the article illustrates that a high level of cartographic quality for PFLP can be achieved through the integrated approach, comparable to the lettering produced by an expert cartographer.

A lagrangean decomposition for the maximum independent set problem applied to map labeling

The Maximum Independent Set (MIS) problem is a well-known problem where the aim is to find the maximum cardinality independent set in an associated graph. Map labeling problems can often be modeled as a MIS problem in a conflict graph, where labels are selected to be placed near graphical features not allowing overlaps (conflicts) between labels or between labels and features. However, the MIS problem is NP-hard and exact techniques present difficulties for solving some instances. Thus, this paper presents a Lagrangean decomposition to solve a map labeling problem, the Point-Feature Label Placement problem. We treated the problem by a conflict graph that is partitioned into small sub-problems and copies of some decision variables are done. These copies are used in the sub-problems constraints and in some constraints to ensure the equality between the original variables and their copies. After these steps, we relax these copy constraints in a Lagrangean way. Using instances proposed in the literature, our approach was able to prove the optimality for all of them, except one, and the results were better than the ones provided by a commercial solver.

POPMUSIC for the point feature label placement problem

Point feature label placement is the problem of placing text labels adjacent to point features on a map so as to maximize legibility. The goal is to choose positions for the labels that do not give rise to label overlaps and that minimize obscuration of features. A practical goal is to minimize the number of overlaps while considering cartographic preferences. This article proposes a new heuristic for solving the point feature label placement problem based on the application of the POPMUSIC frame. Computational experiments show that the proposed heuristic outperformed other recent metaheuristics approaches in the literature. Experiments with problem instances involving up to 10 million points show that the computational time of the proposed heuristic increases almost linearly with the problem size. New problem instances based on real data with more than 13,000 labels are proposed.

Particle-based labeling: Fast point-feature labeling without obscuring other visual features

In many information visualization techniques, labels are an essential part to communicate the visualized data. To preserve the expressiveness of the visual representation, a placed label should neither occlude other labels nor visual representatives (e.g., icons, lines) that communicate crucial information. Optimal, non-overlapping labeling is an NP-hard problem. Thus, only a few approaches achieve a fast non-overlapping labeling in highly interactive scenarios like information visualization. These approaches generally target the point-feature label placement (PFLP) problem, solving only label-label conflicts. This paper presents a new, fast, solid and flexible 2D labeling approach for the PFLP problem that additionally respects other visual elements and the visual extent of labeled features. The results (number of placed labels, processing time) of our particle-based method compare favorably to those of existing techniques. Although the esthetic quality of non-real-time approaches may not be achieved with our method, it complies with practical demands and thus supports the interactive exploration of information spaces. In contrast to the known adjacent techniques, the flexibility of our technique enables labeling of dense point clouds by the use of non-occluding distant labels. Our approach is independent of the underlying visualization technique, which enables us to demonstrate the application of our labeling method within different information visualization scenarios.

Fast Point-Feature Label Placement for Dynamic Visualizations

Fast Point-Feature Label Placement for Dynamic Visualizations

Fast Point-Feature Label Placement for Dynamic Visualizations

This paper describes a fast approach to automatic point label de-confliction on interactive maps. The general Map Labeling problem is NP-hard and has been the subject of much study for decades. Computerized maps have introduced interactive zooming and panning, which has intensified the problem. Providing dynamic labels for such maps typically requires a time-consuming pre-processing phase. In the realm of visual analytics, however, the labeling of interactive maps is further complicated by the use of massive datasets laid out in arbitrary configurations, thus rendering reliance on a pre-processing phase untenable. This paper offers a method for labeling point-features on dynamic maps in real time without pre-processing. The algorithm presented is efficient, scalable, and exceptionally fast; it can label interactive charts and diagrams at speeds of multiple frames per second on maps with tens of thousands of nodes. To accomplish this, the algorithm employs a novel geometric de-confliction approach, the ‘trellis strategy,’ along with a unique label candidate cost analysis to determine the “least expensive” label configuration. The speed and scalability of this approach make it well-suited for visual analytic applications.

Placing Text Labels on Maps and Diagrams using Genetic Algorithms with Masking

Cartographic label placement is one of the most time-consuming tasks in the production of high quality maps and other high quality graphical displays. It is essential that text labels used to identify various features and objects be placed in a clear and unobscured manner. In this article we are concerned with the placement of labels for point features. Specifically, the point feature label placement (PFLP) problem is the problem of placing text labels to point features on a map, graph, or diagram in such a manner so as to maximize legibility. The PFLP problem has been shown to be NP-hard. We propose a heuristic method for the PFLP problem based on genetic algorithms (GA), an adaptive, robust, search and optimization technique based on the principles of natural genetics and survival of the fittest. In particular we emphasize the notion of masking to preserve optimal subsequences in chromosomes and prevent their disruption during crossover and mutation. We ran our algorithms on randomly placed point features in a region, and on datasets from various regions of the USA map with great success. Our GA Implementation with masking solved each of the test cases extremely well, and proved to be an excellent heuristic for solving the PFLP problem. Furthermore, our GA with masking performed significantly better than other PFLP algorithms from the literature.

Practical Results Using Simulated Annealing for Point Feature Label Placement

This paper describes a production-oriented computer program for point feature label placement on petroleum industry basemaps. The program uses a simulated annealing algorithm to find a feasible and near optimal solution to the problem of placing point feature labels on a map near their corresponding symbols, without overplotting each other or other point feature symbols. The program implementation extends recently published work on the use of stochastic algorithms to solve the label placement problem. Results show that simulated annealing provides an efficient and robust solution to map labeling problems for the petroleum industry.

An empirical study of algorithms for point-feature label placement

A major factor affecting the clarity of graphical displays that include text labels is the degree to which labels obscure display features (including other labels) as a result of spatial overlap. Point-feature label placement (PFLP) is the problem of placing text labels adjacent to point features on a map or diagram so as to maximize legibility. This problem occurs frequently in the production of many types of informational graphics, though it arises most often in automated cartography. In this paper we present a comprehensive treatment of the PFLP problem, viewed as a type of combinatorial optimization problem. Complexity analysis reveals that the basic PFLP problem and most interesting variants of it are NP-hard. These negative results help inform a survey of previously reported algorithms for PFLP; not surprisingly, all such algorithms either have exponential time complexity or are incomplete. To solve the PFLP problem in practice, then, we must rely on good heuristic methods. We propose two new methods, one based on a discrete form of gradient descent, the other on simulated annealing, and report on a series of empirical tests comparing these and the other known algorithms for the problem. Based on this study, the first to be conducted, we identify the best approaches as a function of available computation time.