Interactive Selection:
(1)Multiple output
(2)Classified output
- only search in a specified category
- search those picture with some special features: ex: painting of human being
(to be continued..)
2007年4月22日
2007年4月20日
[imPainter] Problems & Approach
Problems to be solved:
(1) Input analysis
(2) Output database construction
(3) Link between input and output
(1) Input analysis
Basic:
-Color/Intensity
-Edge
Advance:
-Background substraction
-Object recognition
-Texture analysis
(2) Output database construction
-Same analysis rules with input
-Different approach (supervised classification)
(3) Link between input & output
(to be continued..)
(1) Input analysis
(2) Output database construction
(3) Link between input and output
(1) Input analysis
Basic:
-Color/Intensity
-Edge
Advance:
-Background substraction
-Object recognition
-Texture analysis
(2) Output database construction
-Same analysis rules with input
-Different approach (supervised classification)
(3) Link between input & output
(to be continued..)
[imPainter] Related works
Philips Flat TV
It provides two features that adapt to the context:
(1) Each pixel of the incoming picture is enhanced to better match the surrounding pixels, resulting in a more natural picture. Artefacts in compressed multimedia content are detected and reduced, ensuring that the picture is clean and razor sharp. Digital noise reduction ensures that the picture is perfectly smooth yet razor sharp.
(2)By providing ambient light to complement the colors and light intensity of the onscreen image, it adds a new dimension to the viewing experience, completely immersing users into the content they are watching.
Color Hamonization (SIGGRAPH 2006)
Harmonic colors are sets of colors that are aesthetically pleasing in terms of human visual perception. In this paper,a method that enhances the harmony among the colors of a given photograph or of a general image was presented, while remaining faithful, as much as possible, to the original colors. It then allows a graceful shifting of hue values so as to fit the harmonic scheme while considering spatial coherence among colors of neighboring pixels using an optimization technique.
Fast Multiresolution Image Querying (SIGGRAPH 1995)
Searching in an image database using aquery image that is similar to the intended target. The query imagemay be a hand-drawn sketch or a (potentially low-quality) scan of the image to be retrieved. The searching algorithm makes use of multiresolution wavelet decompositions of the query and databaseimages. The coefficients of these decompositions are distilled intosmall "signatures" for each image. This metric essentiallycompares how many significant wavelet coefficients the query has incommon with potential targets.
Application: retrievr
(to be continued....)
It provides two features that adapt to the context:
(1) Each pixel of the incoming picture is enhanced to better match the surrounding pixels, resulting in a more natural picture. Artefacts in compressed multimedia content are detected and reduced, ensuring that the picture is clean and razor sharp. Digital noise reduction ensures that the picture is perfectly smooth yet razor sharp.
(2)By providing ambient light to complement the colors and light intensity of the onscreen image, it adds a new dimension to the viewing experience, completely immersing users into the content they are watching.
Color Hamonization (SIGGRAPH 2006)
Harmonic colors are sets of colors that are aesthetically pleasing in terms of human visual perception. In this paper,a method that enhances the harmony among the colors of a given photograph or of a general image was presented, while remaining faithful, as much as possible, to the original colors. It then allows a graceful shifting of hue values so as to fit the harmonic scheme while considering spatial coherence among colors of neighboring pixels using an optimization technique.
Fast Multiresolution Image Querying (SIGGRAPH 1995)
Searching in an image database using aquery image that is similar to the intended target. The query imagemay be a hand-drawn sketch or a (potentially low-quality) scan of the image to be retrieved. The searching algorithm makes use of multiresolution wavelet decompositions of the query and databaseimages. The coefficients of these decompositions are distilled intosmall "signatures" for each image. This metric essentiallycompares how many significant wavelet coefficients the query has incommon with potential targets.
Application: retrievr
(to be continued....)
[imPainter] Concept
The original idea was proposed by dAb, and further discussed by Allen and Ting of IVLab. Inspired by the Philips Flat TV, which is designed "around you" by automatically adjusting to the context, we thought making a more intelligent environment adaptive picture frame might be an interesting work.
How it works?
In our design, imPainter is composed by a display and 1-2 cameras on it. When the user brings imPainter to a new environment, the cameras start to capture photos from the surroundings, then we analyse these photos, including colors, lights, diversity, space and even more complicated work such as obejct recognition. After describing the context by some feature vectors, imPainter tries to find one or a set of appropriate pictures (paintings) for this environment.
The Name?
"imPainter" is the name of this work (may change though :P). My research area is HCI, and currently join in another project called "imTop" (an interactive tableTop). In these "im-" projects, we want to make "I" the subject of technology - computer needs to know what people intend to see and get, to make things simple and bring fun to people's everyday life.
From another point of view, "im-" means that computer itself could be anything, just as in this project, it becomes a painter.
The name "im-" is also an abbreviation for some specific meanings: "imTop" is for "Interactive Multi-resolution tableTop", and "imPainter" is for "Intelligent Ambience Painter".
How it works?
In our design, imPainter is composed by a display and 1-2 cameras on it. When the user brings imPainter to a new environment, the cameras start to capture photos from the surroundings, then we analyse these photos, including colors, lights, diversity, space and even more complicated work such as obejct recognition. After describing the context by some feature vectors, imPainter tries to find one or a set of appropriate pictures (paintings) for this environment.
The Name?
"imPainter" is the name of this work (may change though :P). My research area is HCI, and currently join in another project called "imTop" (an interactive tableTop). In these "im-" projects, we want to make "I" the subject of technology - computer needs to know what people intend to see and get, to make things simple and bring fun to people's everyday life.
From another point of view, "im-" means that computer itself could be anything, just as in this project, it becomes a painter.
The name "im-" is also an abbreviation for some specific meanings: "imTop" is for "Interactive Multi-resolution tableTop", and "imPainter" is for "Intelligent Ambience Painter".
Project "imPainter"
"imPainter" is one work of the "im-" projects, and is also the term project topic for the course "Multimedia Analysis and Indexing", 2007 Spring, National Taiwan University.
All posts about this topic will have title starting with [imPainter] and labeled as "impainter".
All posts about this topic will have title starting with [imPainter] and labeled as "impainter".
訂閱：
文章 (Atom)
