摘要:
An edge extracting unit of a contour image generator generates an edge image of an input image using an edge extraction filter, etc. A foreground processing unit extracts the foreground from the input image using a background image and expands the foreground to generate an expanded foreground image. The foreground processing unit further generates a foreground boundary image constructed of the boundary of the expanded foreground region. A mask unit masks the edge image using the expanded foreground image to eliminate edges in the background. A synthesis unit synthesizes the masked edge image and the foreground boundary image to generate a contour image.
摘要:
A tracking start-and-stop determining unit of a tracking processing unit determines whether to start or stop tracking based on a predetermined condition. During tracking, a sampling unit 29 creates or eliminates particles using a probability density distribution estimated for the previous image frame. A first to nth model tracking units transit the particles according to motion models respectively assigned, and observe the likelihood of a candidate curve corresponding to each particle with respect to an edge image generated by an image processing unit. A tracking integration unit determines a motion model which causes the highest likelihood, and estimates a probability density distribution using an observation likelihood based on the motion model.
摘要:
Candidate contour curves for a tracking object in the current frame are determined using a particle filter, based on the existence probability distribution of the tracking object in a frame which is one frame previous to the current frame. To match a candidate curve against a contour image of the current frame, a processing to search for the closest contour to the candidate curves is divided for each knot constituting the candidate contour curve and is executed in parallel by a plurality of processors. Each image data on a search region for each knot to be processed are copied from a contour image stored in an image storage to the respective local memories.
摘要:
A positional data acquisition unit of an action detector acquires positional data indicating the position of an image of a light-emitting part of a light-emitting device held by a user in an image frame at each time step, and also acquires curve data for the head contour at each time step estimated as a result of visual tracking by a tracking processor. A history storage unit stores a history of the positional data for an image of a light-emitting part and the curved data for the head contour. A determination criteria storage unit stores the criteria for determining that a predefined action is performed by referring to the time-dependent change in the relative position of the image of the light-emitting part in relation to the curve representing the head contour. An action determination unit determines whether the action is performed based on the actual data.
摘要:
A face detection processing unit performs a face detection process by rotating an image in increments of a predetermined angle and acquires a rotation angle at which a face is detected. An angle correction unit acquires an angle between the face and a shoulder by pattern matching and corrects the rotation angle of the image. A human-image orientation identification unit identifies a correct orientation of a human image based on the rotation angle. An image-characteristic analysis unit analyzes a frequency distribution and a brightness distribution of a non-human image. A non-human image orientation identification unit identifies the correct orientation of a non-human image based on distribution characteristics of a frequency of brightness with respect to an axis in a predetermined direction. An image-data updating unit incorporates information regarding the correct orientation in the image data.
摘要:
A photograph classification unit classifies the photographs for each scene in a music-feature information determination unit. A photograph feature acquisition unit identifies a feature of the photograph based on additional information of the photograph and a result of face recognition of the image. A tempo determination unit determines a tempo of music based on a time zone of image capturing, a range of the number of people captured in a photograph, the degree of smile, activities, etc. A melody determination unit determines information that confines a title, a feature value, a genre, etc., based on an event, a time zone, and a season at the time of image capturing, a city and a country in which the image capturing has occurred, etc. Based on the determined feature of the music, a music data output unit extracts and then presents matching music.
摘要:
To provide an operation input device capable of realizing operation input by changing the posture of the operation input device even without a posture determining sensor. This is an operation input device for use by a user while being held by his/her hand, for obtaining a captured image captured by an image capturing unit provided to the operation input device, and determining the posture of the operation input device, based on the result of analysis on the captured image obtained.
摘要:
Candidate contour curves for a tracking object in the current frame are determined using a particle filter, based on the existence probability distribution of the tracking object in a frame which is one frame previous to the current frame. To match a candidate curve against a contour image of the current frame, a processing to search for the closest contour to the candidate curves is divided for each knot constituting the candidate contour curve and is executed in parallel by a plurality of processors. Each image data on a search region for each knot to be processed are copied from a contour image stored in an image storage to the respective local memories.
摘要:
In a position history analysis unit for analyzing a position history to acquire movement history information such as places of stay, a speed history computation unit refers to the position history to determine a speed occurring at each point of time and to obtain a speed history. A categorization unit organizes a time axis in the position history, and, ultimately, data for position coordinates occurring at respective points of time, into segments of stay and segments of movement, by checking the speed history against a threshold value. A merge unit detects segments of stay which occur at different points of time but which are identified as indicating stays in the same place, by feeding the time flow, categorized into segments of stay and segments of movement by the categorization unit based on the speed, back to the position coordinates in the respective segments. The merge unit merges the segments of stay as needed. An information generation unit evaluates a set of position coordinates and speed values in each segment of stay and in each segment of movement thus obtained and derives predetermined information on each segment.
摘要:
A photograph classification unit classifies the photographs for each scene in a music-feature information determination unit. A photograph feature acquisition unit identifies a feature of the photograph based on additional information of the photograph and a result of face recognition of the image. A tempo determination unit determines a tempo of music based on a time zone of image capturing, a range of the number of people captured in a photograph, the degree of smile, activities, etc. A melody determination unit determines information that confines a title, a feature value, a genre, etc., based on an event, a time zone, and a season at the time of image capturing, a city and a country in which the image capturing has occurred, etc. Based on the determined feature of the music, a music data output unit extracts and then presents matching music.