Using Fuzzy Hybrid Features to Classify Strokes in Interactive Sketches

Abstract

A novel method is presented based on fuzzy hybrid-based features to classify strokes into 2D line drawings, and a human computer interactive system is developed for assisting designers in conceptual design stage. Fuzzy classifiers are built based on some geometric features and speed features. The prototype system can support rapid classification based on fuzzy classifiers, and the classified stroke is then fitted with a 2D geometry primitive which could be a line segment, polyline, circle, circular arc, ellipse, elliptical arc, hyperbola, and parabola. The human computer interaction can determine the ambiguous results and then revise the misrecognitions. The test results showed that the proposed method can support online freehand sketching based on conceptual design with no limitation on drawing sequence and direction while achieving a satisfactory interpretation rate.

Tolerance Zone-Based Grouping Method for Online Multiple Overtracing Freehand Sketches

“…A stroke group from the grouping process may contain one or more geometric primitives. Before the fitting process, a fuzzy classification method in [11] is used to recognize each stroke of the group as a line segment, a polyline or a conic curve.…”

“…When a group is merely composed of line segment strokes, it would be fitted into a line segment if the whole group is recognized as a line segment by comparing the length and width of its bounding box [11]. A line segment is obtained from connecting the two furthest polygonal approximate points of the group.…”

“…Currently, in the field of reconstruction of 3D models from 2D sketches, sketch interpretation is usually based on primitives. Such methods addresses the uniqueness problem of the sketch interpretation, which involves sketch segmentation [6,8], stroke grouping [9,10], line fitting [11], endpoint fusion, and 2D regularity enhancement [12].…”

“…e work presented here provides a foundation for further line drawing reconstruction. In FSR-MOS system, our previous fuzzy classification method [11] is used to recognize primitives as straight lines, polylines, ellipses, elliptical arcs, circles, circular arcs, parabolas, or hyperbolas. Among them, ellipses, elliptical arcs, circles, circular arcs, parabolas, and hyperbolas are collectively called conic curves in this paper.…”

Optimasi Fungsi Keanggotaan Fuzzy Berbasis Algoritma Modified Particle Swarm Optimization

“…Fungsi keanggotaan merupakan inti dari disain pengambilan keputusan dengan logika fuzzy [1]. Hal senada juga dikemukakan oleh [2] bahwa pemilihan fungsi keanggotaan merupakan kunci utama pada logika fuzzy karena disain fungsi keanggotaan sangat mempengaruhi konsistensi, ketepatan dan kualitas keputusan yang diselesaikan pada sistem tersebut.…”