We demonstrate an intuitive and interactive prototype sketching system that reconstructs 3D polyh... more We demonstrate an intuitive and interactive prototype sketching system that reconstructs 3D polyhedrons incrementally from 2D freehand sketch. The system avoids conventional menu interface and provides an interface mimicking pencil and paper that takes 2D freehand sketch as its direct input. The system offers a natural sketching interface by allowing sketching with discontinuous, overlapping and multiple strokes. The input sketch is a natural line drawing with hidden line removed that depicts a 3D object in real world. The system is configured to interpret sketch under isometric view, which is the most popular viewpoint among designers. The system automatically tidies-up the freehand 2D sketch and reconstructs it into a 3D object. The object can then be transformed, which provides immediate feedback to the designers. New sketch can be drawn over the 3D object, and then be reconstructed into 3D by referring to the existing 3D object from the current viewpoint. This method allows the designers add new parts on the ``backside'' of the initial sketch. The incremental modelling from sketch enables easy reconstruction of complex 3D objects comparing with the current 3D reconstruction algorithms. It also enables a more detailed object to be reconstructed.
This paper presents an online personalised non-photorealistic rendering (NPR) technique for 3D mo... more This paper presents an online personalised non-photorealistic rendering (NPR) technique for 3D models generated from interactively sketched input. This technique has been integrated into a sketch-based modelling system. It lets users interact with computers by drawing naturally, without specifying the number, order, or direction of strokes. After sketches are interpreted as 3D objects, they can be rendered with personalised drawing styles so that the reconstructed 3D model can be presented in a sketchy style similar in appearance to what have been drawn for the 3D model. This technique captures the user's drawing style without using template or prior knowledge of the sketching style. The personalised rendering style can be applied to both visible and initially invisible geometry. The rendering strokes are intelligently selected from the input sketches and mapped to edges of the 3D object. In addition, non-geometric information such as surface textures can be added to the recognised object in different sketching modes. This will integrate sketch-based incremental 3D modelling and NPR into conceptual design.
Recently, the generalization framework in coevolutionary learning has been theoretically formulat... more Recently, the generalization framework in coevolutionary learning has been theoretically formulated and demonstrated in the context of game-playing. Generalization performance of a strategy (solution) is estimated using a collection of random test strategies (test cases) by taking the average game outcomes, with confidence bounds provided by Chebyshev's theorem. Chebyshev's bounds have the advantage that they hold for any distribution of game outcomes. However, such a distributionfree framework leads to unnecessarily loose confidence bounds. In this paper, we have taken advantage of the near-Gaussian nature of average game outcomes and provided tighter bounds based on parametric testing. This enables us to use small samples of test strategies to guide and improve the co-evolutionary search. We demonstrate our approach in a series of empirical studies involving the iterated prisoner's dilemma (IPD) and the more complex Othello game in a competitive co-evolutionary learning setting. The new approach is shown to improve on the classical co-evolutionary learning in that we obtain increasingly higher generalization performance using relatively small samples of test strategies. This is achieved without large performance fluctuations typical of the classical approach. The new approach also leads to faster co-evolutionary search where we can strictly control the condition (sample sizes) under which the speedup is achieved (not at the cost of weakening precision in the estimates).
We present an intuitive and interactive freehand sketching interface for 3D polyhedrons reconstru... more We present an intuitive and interactive freehand sketching interface for 3D polyhedrons reconstruction. The interface mimics sketching with pencil on paper and takes freehand sketches as input directly. The sketching environment is natural by allowing sketching with discontinuous, overlapping and multiple strokes. The input sketch is a natural line drawing with hidden lines removed that depicts a 3D object in an isometric view. The line drawing is interpreted by a series of 2D tidy-up processes to produce a vertex-edge graph for 3D reconstruction. A novel reconstruction approach based on three-line-junction analysis and planarity constraint is then used to approximate the 3D geometry and topology of the graph. The reconstructed object can be transformed so that it can be viewed from different viewpoints for interactive design or as immediate feedback to the designers. A new sketch can then be added to the existing 3D object, and reconstructed into 3D by referring to the existing 3D object from the current viewpoint. The incremental modeling enables a 3D object to be reconstructed from multiple sketching sessions from different viewpoints. However, the interface is limited to reconstructing trihedrons from sketches without T-junctions to avoid ambiguity in the hidden topology determination.
This paper presents a novel method for interpreting overtracing freehand sketch. The overtracing ... more This paper presents a novel method for interpreting overtracing freehand sketch. The overtracing strokes are interpreted as sketch content and are used to generate 2D geometric primitives. The approach consists of four stages: stroke classification, strokes grouping and fitting, 2D tidy-up with endpoint clustering and parallelism correction, and in-context interpretation. Strokes are first classified into lines and curves by a linearity test. It is followed by an innovative strokes grouping process that handles lines and curves separately. The grouped strokes are fitted with 2D geometry and further tidied-up with endpoint clustering and parallelism correction. Finally, the in-context interpretation is applied to detect incorrect stroke interpretation based on geometry constraints and to suggest a most plausible correction based on the overall sketch context. The interpretation ensures sketched strokes to be interpreted into meaningful output. The interface overcomes the limitation where only a single line drawing can be sketched out as in most existing sketching programs, meanwhile is more intuitive to the user.
We demonstrate an intuitive and interactive prototype sketching system that reconstructs 3D polyh... more We demonstrate an intuitive and interactive prototype sketching system that reconstructs 3D polyhedrons incrementally from 2D freehand sketch. The system avoids conventional menu interface and provides an interface mimicking pencil and paper that takes 2D freehand sketch as its direct input. The system offers a natural sketching interface by allowing sketching with discontinuous, overlapping and multiple strokes. The input sketch is a natural line drawing with hidden line removed that depicts a 3D object in real world. The system is configured to interpret sketch under isometric view, which is the most popular viewpoint among designers. The system automatically tidies-up the freehand 2D sketch and reconstructs it into a 3D object. The object can then be transformed, which provides immediate feedback to the designers. New sketch can be drawn over the 3D object, and then be reconstructed into 3D by referring to the existing 3D object from the current viewpoint. This method allows the designers add new parts on the ``backside'' of the initial sketch. The incremental modelling from sketch enables easy reconstruction of complex 3D objects comparing with the current 3D reconstruction algorithms. It also enables a more detailed object to be reconstructed.
This paper presents an online personalised non-photorealistic rendering (NPR) technique for 3D mo... more This paper presents an online personalised non-photorealistic rendering (NPR) technique for 3D models generated from interactively sketched input. This technique has been integrated into a sketch-based modelling system. It lets users interact with computers by drawing naturally, without specifying the number, order, or direction of strokes. After sketches are interpreted as 3D objects, they can be rendered with personalised drawing styles so that the reconstructed 3D model can be presented in a sketchy style similar in appearance to what have been drawn for the 3D model. This technique captures the user's drawing style without using template or prior knowledge of the sketching style. The personalised rendering style can be applied to both visible and initially invisible geometry. The rendering strokes are intelligently selected from the input sketches and mapped to edges of the 3D object. In addition, non-geometric information such as surface textures can be added to the recognised object in different sketching modes. This will integrate sketch-based incremental 3D modelling and NPR into conceptual design.
Recently, the generalization framework in coevolutionary learning has been theoretically formulat... more Recently, the generalization framework in coevolutionary learning has been theoretically formulated and demonstrated in the context of game-playing. Generalization performance of a strategy (solution) is estimated using a collection of random test strategies (test cases) by taking the average game outcomes, with confidence bounds provided by Chebyshev's theorem. Chebyshev's bounds have the advantage that they hold for any distribution of game outcomes. However, such a distributionfree framework leads to unnecessarily loose confidence bounds. In this paper, we have taken advantage of the near-Gaussian nature of average game outcomes and provided tighter bounds based on parametric testing. This enables us to use small samples of test strategies to guide and improve the co-evolutionary search. We demonstrate our approach in a series of empirical studies involving the iterated prisoner's dilemma (IPD) and the more complex Othello game in a competitive co-evolutionary learning setting. The new approach is shown to improve on the classical co-evolutionary learning in that we obtain increasingly higher generalization performance using relatively small samples of test strategies. This is achieved without large performance fluctuations typical of the classical approach. The new approach also leads to faster co-evolutionary search where we can strictly control the condition (sample sizes) under which the speedup is achieved (not at the cost of weakening precision in the estimates).
We present an intuitive and interactive freehand sketching interface for 3D polyhedrons reconstru... more We present an intuitive and interactive freehand sketching interface for 3D polyhedrons reconstruction. The interface mimics sketching with pencil on paper and takes freehand sketches as input directly. The sketching environment is natural by allowing sketching with discontinuous, overlapping and multiple strokes. The input sketch is a natural line drawing with hidden lines removed that depicts a 3D object in an isometric view. The line drawing is interpreted by a series of 2D tidy-up processes to produce a vertex-edge graph for 3D reconstruction. A novel reconstruction approach based on three-line-junction analysis and planarity constraint is then used to approximate the 3D geometry and topology of the graph. The reconstructed object can be transformed so that it can be viewed from different viewpoints for interactive design or as immediate feedback to the designers. A new sketch can then be added to the existing 3D object, and reconstructed into 3D by referring to the existing 3D object from the current viewpoint. The incremental modeling enables a 3D object to be reconstructed from multiple sketching sessions from different viewpoints. However, the interface is limited to reconstructing trihedrons from sketches without T-junctions to avoid ambiguity in the hidden topology determination.
This paper presents a novel method for interpreting overtracing freehand sketch. The overtracing ... more This paper presents a novel method for interpreting overtracing freehand sketch. The overtracing strokes are interpreted as sketch content and are used to generate 2D geometric primitives. The approach consists of four stages: stroke classification, strokes grouping and fitting, 2D tidy-up with endpoint clustering and parallelism correction, and in-context interpretation. Strokes are first classified into lines and curves by a linearity test. It is followed by an innovative strokes grouping process that handles lines and curves separately. The grouped strokes are fitted with 2D geometry and further tidied-up with endpoint clustering and parallelism correction. Finally, the in-context interpretation is applied to detect incorrect stroke interpretation based on geometry constraints and to suggest a most plausible correction based on the overall sketch context. The interpretation ensures sketched strokes to be interpreted into meaningful output. The interface overcomes the limitation where only a single line drawing can be sketched out as in most existing sketching programs, meanwhile is more intuitive to the user.
Uploads
Papers by Daychyi Ku