I like deriving inspiration from everyday people and nature to find out new problems and their solutions. Observing other people perform tasks results in some unique self-discovery about how humans learn, adapt and apply their mind to find solutions to problems.
My broad research aim is to bring together knowledge from various professions and subjects to create new innovative products for tomorrow's intellectual, professional and common man. My experience is that many times the wheel is either re-invented or its invention is delayed because of lack of knowledge existing in other fields of science and technology. Establishing connections between diverse fields will open up a plethora of opportunities for innovation.
Sketch-based Geometry Reconstruction |
||
|
|
The aim of this project, entitled SMARTPAPER, is to convert 2D rough, hand-drawn sketches of solid objects into 3D geometry, using minimal and intuitive user interaction. We concentrate on developing core algorithms for the interactive reconstruction of 3D geometry without any prior knowledge about the nature of solid objects being sketched. Using SMARTPAPER on a Tablet PC, the user can sketch exactly as on paper, and even use intuitive gestures to aid in the reconstruction process. Sketch-based reconstruction of 3D geometry finds applications in the early stages of mechanical engineering and architectural design. Publications:
|
|
||
|
|
|
|
Sketch-assisted Photo Navigation |
||
|
|
The aim of this project, entitled Peek-in-the-Pic, is to impart a third dimension to photographs, so that it is possible to fly through them. Our system offers an easy interface to interactively construct 3D geometry starting from a single, casually taken photograph of architectural objects like buildings. Using our system, the user can virtually ``fly through'' a photograph to get a feeling of ``being there'', even without having actually visited the place (virtual photo tourism). Publications:
|
|
||
|
|
|
|
Sketch-based Inverse Lighting |
||
|
|
The aim of this project, entitled Crayon Lighting, is to provide a sketch-based interaction metaphor for a completely goal-based way of designing lights for a particular scene. A user roughly sketches desired lighting effects like highlights and shadows in a given scene, from which the system determines where to place lights to produce these effects. Such a system finds applications in designing lights for buildings and in animations to produce desired cinematic effects. Publications: |
|
||
|
|
|
|
Rendering line models with fidelity |
||
|
|
The style and abstract nature of hand-drawn sketches is important in many design and artistic applications, due to which it is undesirable to refine, beautify or even parse them. The aim of this project is to retain the original nature of hand-drawn sketches and concentrate on rendering them automatically and faithfully. We address several fundamental challenges of rendering 2D/3D wire-frame sketches, propose and implement a rendering pipeline based on lines and curves as basic rendering primitives. Not only does this find important applications in sketch-based art and design, but also in more efficient data representations for non-photorealistic rendering of geometry. Publications:
|
|
||
These projects are not a major part of my Ph.D. dissertation. These projects are collaborative in nature, combining my work with other work done in our research group.
Sketch-based Segmentation of Models |
|
|
When modeling with scanned outdoor models, being able to select a subset of the points efficiently that collectively represent an object is an important operation. We propose a sketch-based interface that allows such segmentation. The user marks object and background regions by placing strokes, and the tool segments out the marked object(s). To refine the results, the user simply moves the camera to a different location and repeats the process. Our method works for general 3D point models and not just range images. Publications:
|
|
|
|
|
|
Hardware-assisted Motion Blur Generation |
|
|
The aim of this project is to interactively generate motion blur effects for animated 3D scenes, using commodity graphics hardware. We have developed a method to generate motion blur by interactive Line Integral Convolution, for Nvidia GeForce4 graphics cards. Publications:
|
|
|
|