Article Free AccessVisual models of plants interacting with their environment Share on Authors: Radomír Měch Department Computer Science, University Calgary, Alberta, Canada T2N 1N4 1N4View Profile , Przemyslaw Prusinkiewicz Authors Info & Claims SIGGRAPH '96: Proceedings the 23rd annual conference graphics and interactive techniquesAugust 1996 Pages 397–410https://doi.org/10.1145/237170.237279Online:01 August 1996Publication History 299citation3,084DownloadsMetricsTotal Citations299Total...

We propose a highly efficient, yet powerful, salient object detection method based on the Minimum Barrier Distance (MBD) Transform. The MBD transform is robust to pixel-value fluctuation, and thus can be effectively applied raw pixels without region abstraction. present an approximate algorithm with 100X speedup over exact algorithm. An error bound analysis also provided. Powered by this fast algorithm, proposed runs at 80 FPS, significantly outperforms previous methods similar speed four...

This paper investigates problems of image style, aesthetics, and quality estimation, which require fine-grained details from high-resolution images, utilizing deep neural network training approach. Existing convolutional networks mostly extracted one patch such as a down-sized crop each example. However, may not always well represent the entire image, cause ambiguity during training. We propose multi-patch aggregation approach, allows us to train models using multiple patches generated...

The use of 3D printing has rapidly expanded in the past couple years. It is now possible to produce 3D-printed objects with exceptionally high fidelity and precision. However, although quality improved, both time print material costs have remained high. Moreover, there no guarantee that a printed model structurally sound. product often does not survive cleaning, transportation, or handling, it may even collapse under its own weight. We present system addresses this issue by providing...

Reconstructing 3D shapes from single-view images has been a long-standing research problem. In this paper, we present DISN, Deep Implicit Surface Network which can generate high-quality detail-rich mesh an 2D image by predicting the underlying signed distance fields. addition to utilizing global features, DISN predicts projected location for each point on image, and extracts local features feature maps. Combining significantly improves accuracy of field prediction, especially areas. To best...

Procedural representations provide powerful means for generating complex geometric structures. They are also notoriously difficult to control. In this article, we present an algorithm controlling grammar-based procedural models. Given a grammar and high-level specification of the desired production, computes production from that conforms specification. This is generated by optimizing over space possible productions grammar. The supports specifications many forms, including shapes analytical...

Article Realistic modeling and rendering of plant ecosystems Share on Authors: Oliver Deussen Otto-von-Guericke Univ. Magdeburg MagdeburgView Profile , Pat Hanrahan Stanford Univ., Stanford, CA CAView Bernd Lintermann ZKM Center for Art Media Karlsruhe KarlsruheView Radomír Měch Calgary, Alta., Matt Pharr Przemyslaw Prusinkiewicz Authors Info & Claims SIGGRAPH '98: Proceedings the 25th annual conference Computer graphics interactive techniquesJuly 1998 Pages...

The paper extends Lindenmayer systems in a manner suitable for simulating the interaction between developing plant and its environment. formalism is illustrated by modeling response of trees to pruning, which yields synthetic images sculptured plants found topiary gardens.

We present a method for generating realistic models of temperate-climate trees and shrubs. This is based on the biological hypothesis that form developing tree emerges from self-organizing process dominated by competition buds branches light or space, regulated internal signaling mechanisms. Simulations this robustly generate wide range bushes. The generated forms can be controlled with variety interactive techniques, including procedural brushes, sketching, editing operations such as...

Abstract Procedural tree models have been popular in computer graphics for their ability to generate a variety of output trees from set input parameters and simulate plant interaction with the environment realistic placement virtual scenes. However, defining such is difficult task. We propose an inverse modelling approach stochastic that takes polygonal as estimates procedural model so it produces similar input. Our framework based on novel parametric generation uses Monte Carlo Markov...

Recently, researchers have developed black-box approaches to mine design and interaction data from mobile apps. Although the captured during this mining is descriptive, it does not expose semantics of UIs: what elements on screen mean how they are used. This paper introduces an automatic approach for generating semantic annotations app UIs. Through iterative open coding 73k UI 720 screens, we contribute a lexical database 25 types components, 197 text button concepts, 135 icon classes shared...

Applications in virtual and augmented reality create a demand for rapid creation easy access to large sets of 3D models. An effective way address this is edit or deform existing models based on reference, e.g., 2D image which very acquire. Given such source model target can be image, model, point cloud acquired as depth scan, we introduce 3DN, an end-to-end network that deforms the resemble target. Our method infers per-vertex offset displacements while keeping mesh connectivity fixed. We...

Abstract We propose an optimization framework for 3D printing that seeks to save time and the support material required print shapes. Three‐dimensional technology is rapidly maturing may revolutionize how we manufacture objects. The total cost of printing, however, governed by numerous factors which include not only price printer but also amount fabricate shape. Our PackMerger converts input watertight mesh into a shell hollowing its inner parts. then divided segments. location splits...

Automatic image aesthetics rating has received a growing interest with the recent breakthrough in deep learning. Although many studies exist for learning generic or universal model, investigation of models incorporating individual user's preference is quite limited. We address this personalized problem by showing that individual's aesthetic preferences exhibit strong correlations content and attributes, hence deviation perception from predictable. To accommodate our study, we first collect...

Design patterns have proven useful in many creative fields, providing content creators with archetypal, reusable guidelines to leverage projects. Creating such patterns, however, is a time-consuming, manual process, typically relegated few experts any given domain. In this paper, we describe an algorithmic method for learning design directly from data using techniques natural language processing and structured concept learning. Given set of labeled, hierarchical designs as input, induce...

We aim at detecting salient objects in unconstrained images. In images, the number of (if any) varies from image to image, and is not given. present a object detection system that directly outputs compact set windows, if any, for an input image. Our leverages Convolutional-Neural-Network model generate location proposals objects. Location tend be highly overlapping noisy. Based on Maximum Posteriori principle, we propose novel subset optimization framework windows out noisy proposals....

Cropping is one of the most common tasks in image editing for improving aesthetic quality a photograph. In this paper, we propose new, photo cropping system which combines three models: visual composition, boundary simplicity, and content preservation. The composition model measures given crop. Instead manually defining rules or score functions learn from large set well-composed images via discriminative classifier training. simplicity clearness crop to avoid object cutting-through....

Edges, boundaries and contours are important subjects of study in both computer graphics vision. On one hand, they the 2D elements that convey 3D shapes, on other indicative occlusion events thus separation objects or semantic concepts. In this paper, we aim to generate contour drawings, boundary-like drawings capture outline visual scene. Prior art often cast problem as boundary detection. However, set cues presented detection output different from ones also artistic style is ignored. We...

Procedural modeling techniques can produce high quality visual content through complex rule sets. However, controlling the outputs of these for design purposes is often notoriously difficult users due to large number parameters involved in sets and also their non-linear relationship resulting content. To circumvent this problem, we present a sketch-based approach procedural modeling. Given an approximate abstract hand-drawn 2D sketch provided by user, our algorithm automatically computes set...

Abstract We present an important step towards the solution of problem inverse procedural modeling by generating parametric context‐free L‐systems that represent input 2D model. The L‐system rules efficiently code regular structures and parameters properties structure transformations. algorithm takes as a vector image is composed atomic elements, such curves poly‐lines. Similar elements are recognized assigned terminal symbols alphabet. symbols' position orientation pair‐wise compared...

We present a dynamic tree modeling and representation technique that allows complex models to interact with their environment. Our method uses changes in the light distribution proximity solid obstacles other trees as approximations of biologically motivated transformations on skeletal tree's main branches its procedurally generated foliage. Parts are transformed only when required, thus our approach is much faster than common algorithms such Open L-Systems or space colonization methods....

People can immediately and precisely identify that an image contains 1, 2, 3 or 4 items by a simple glance. The phenomenon, known as Subitizing, inspires us to pursue the task of Salient Object Subitizing (SOS), i.e. predicting existence number salient objects in scene using holistic cues. To study this problem, we propose new dataset annotated online crowdsourcing marketplace. We show proposed subitizing technique end-to-end Convolutional Neural Network (CNN) model achieves significantly...