Showing papers on "Object (computer science) published in 2009"

Learning to detect unseen object classes by between-class attribute transfer

Abstract: We study the problem of object classification when training and test classes are disjoint, i.e. no training examples of the target classes are available. This setup has hardly been studied in computer vision research, but it is the rule rather than the exception, because the world contains tens of thousands of different object classes and for only a very few of them image, collections have been formed and annotated with suitable class labels. In this paper, we tackle the problem by introducing attribute-based classification. It performs object detection based on a human-specified high-level description of the target objects instead of training images. The description consists of arbitrary semantic attributes, like shape, color or even geographic information. Because such properties transcend the specific learning task at hand, they can be pre-learned, e.g. from image datasets unrelated to the current task. Afterwards, new classes can be detected based on their attribute representation, without the need for a new training phase. In order to evaluate our method and to facilitate research in this area, we have assembled a new large-scale dataset, “Animals with Attributes”, of over 30,000 animal images that match the 50 classes in Osherson's classic table of how strongly humans associate 85 semantic attributes with animal classes. Our experiments show that by using an attribute layer it is indeed possible to build a learning object detection system that does not require any training images of the target classes.

WhereNext: a location predictor on trajectory pattern mining

Abstract: The pervasiveness of mobile devices and location based services is leading to an increasing volume of mobility data.This side eect provides the opportunity for innovative methods that analyse the behaviors of movements. In this paper we propose WhereNext, which is a method aimed at predicting with a certain level of accuracy the next location of a moving object. The prediction uses previously extracted movement patterns named Trajectory Patterns, which are a concise representation of behaviors of moving objects as sequences of regions frequently visited with a typical travel time. A decision tree, named T-pattern Tree, is built and evaluated with a formal training and test process. The tree is learned from the Trajectory Patterns that hold a certain area and it may be used as a predictor of the next location of a new trajectory finding the best matching path in the tree. Three dierent best matching methods to classify a new moving object are proposed and their impact on the quality of prediction is studied extensively. Using Trajectory Patterns as predictive rules has the following implications: (I) the learning depends on the movement of all available objects in a certain area instead of on the individual history of an object; (II) the prediction tree intrinsically contains the spatio-temporal properties that have emerged from the data and this allows us to define matching methods that striclty depend on the properties of such movements. In addition, we propose a set of other measures, that evaluate a priori the predictive power of a set of Trajectory Patterns. This measures were tuned on a real life case study. Finally, an exhaustive set of experiments and results on the real dataset are presented.

Class-specific Hough forests for object detection

Abstract: We present a method for the detection of instances of an object class, such as cars or pedestrians, in natural images Similarly to some previous works, this is accomplished via generalized Hough transform, where the detections of individual object parts cast probabilistic votes for possible locations of the centroid of the whole object; the detection hypotheses then correspond to the maxima of the Hough image that accumulates the votes from all parts However, whereas the previous methods detect object parts using generative codebooks of part appearances, we take a more discriminative approach to object part detection Towards this end, we train a class-specific Hough forest, which is a random forest that directly maps the image patch appearance to the probabilistic vote about the possible location of the object centroid We demonstrate that Hough forests improve the results of the Hough-transform object detection significantly and achieve state-of-the-art performance for several classes and datasets

Class-Specific Hough Forests for Object Detection

Abstract: We present a method for the detection of instances of an object class, such as cars or pedestrians, in natural images. Similarly to some previous works, this is accomplished via generalized Hough transform, where the detections of individual object parts cast probabilistic votes for possible locations of the centroid of the whole object; the detection hypotheses then correspond to the maxima of the Hough image that accumulates the votes from all parts. However, whereas the previous methods detect object parts using generative codebooks of part appearances, we take a more discriminative approach to object part detection. Towards this end, we train a class-specific Hough forest, which is a random forest that directly maps the image patch appearance to the probabilistic vote about the possible location of the object centroid. We demonstrate that Hough forests improve the results of the Hough-transform object detection significantly and achieve state-of-the-art performance for several classes and datasets.

Combined Vehicle-to-Vehicle Communication and Object Detection Sensing

Abstract: A vehicle awareness system for monitoring remote vehicles relative to a host vehicle. The vehicle awareness system includes at least one object sensing device and a vehicle-to-vehicle communication device. A data collection module is provided for obtaining a sensor object data map and vehicle-to-vehicle object data map. A fusion module merges the sensor object data map and vehicle-to-vehicle object data map for generating a cumulative object data map. A tracking module estimates the relative position of the remote vehicles to the host vehicle.

Data services based on gesture and location information of device

Abstract: With the addition of directional information and gesture based input in a location based services environment, a variety of service(s) can be provided on top of user identification or interaction with specific object(s) of interest. For instance, when a user gestures at or points at a particular item, or gestures at a particular location or place, this creates an opportunity, e.g., an advertising opportunity, for anyone having an interest in that particular item or place to communicate with the user regarding that item or related items at a point in time when the user's focus is on the particular item. User context for the interaction can also be taken into account to supplement the provision of one or more interactive direction based services.

Gesture-based control system for vehicle interfaces

Abstract: Systems and method for controlling vehicle subsystems using gestuies are described. The controlling comprises receiving an image of an object via sensors in a vehicle and outputting gestme data representing a gesture made by the object The object comprises at least one hand and/or finger of a user. The gesture data is an instantaneous state of the object at a point in time in space The controlling comprises automatically detecting the gesture of the object from the gesture data by extracting positions of the object, recovering quantized poses and orientation vectors of the object, and translating the gesture data to a gesture signal. The controlling comprises managing user interactions with the vehicle subsystems in response to the gesture signal

System and method for search portions of objects in images and features thereof

Abstract: Embodiments enable searching of portions of objects in images, including programmatically analyzing each image in a collection in order to determine image data that, for individual images in the collection, represents one or more visual characteristics of a portion of an object shown in that image. A user is enabled to specify one or more search criteria that includes image data, and a search result may be determined based on one or more images in the collection that show a corresponding object that has a portion that satisfies a threshold. The threshold is defined at least in part by the one or more search criteria.

Learning Models for Object Recognition from Natural Language Descriptions

Abstract: We investigate the task of learning models for visual object recognition from natural language descriptions alone. The approach contributes to the recognition of fine-grain object categories, such as animal and plant species, where it may be difficult to collect many images for training, but where textual descriptions of visual attributes are readily available. As an example we tackle recognition of butterfly species, learning models from descriptions in an online nature guide. We propose natural language processing methods for extracting salient visual attributes from these descriptions to use as ‘templates’ for the object categories, and apply vision methods to extract corresponding attributes from test images. A generative model is used to connect textual terms in the learnt templates to visual attributes. We report experiments comparing the performance of humans and the proposed method on a dataset of ten butterfly categories.

Method and apparatus for maintaining the status of objects in computer networks using virtual state machines

Abstract: A network appliance for monitoring, diagnosing and documenting problems among a plurality of devices and processes (objects) coupled to a computer network utilizes periodic polling and collection of object-generated trap data to monitor the status of objects on the computer network. The status of a multitude of objects is maintained in memory utilizing virtual state machines which contain a small amount of persistent data but which are modeled after one of a plurality of finite state machines. The memory further maintains dependency data related to each object which identifies parent/child relationships with other objects at the same or different layers of the OSI network protocol model. A decision engine verifies through on-demand polling that a device is down. A root cause analysis module utilizes status and dependency data to locate the highest object in the parent/child relationship tree that is affected to determine the root cause of a problem. Once a problem has been verified, a “case” is opened and notification alerts may be sent out to one or more devices. A user interface allows all objects within the network to be displayed with their respective status and their respective parent/child dependency objects in various formats.

A cognitive map for object memory in the hippocampus.

Abstract: The hippocampus has been proposed to support a cognitive map, a mental representation of the spatial layout of an environment as well as the nonspatial items encountered in that environment. In the present study, we recorded simultaneously from 43 to 61 hippocampal pyramidal cells as rats performed an object recognition memory task in which novel and repeated objects were encountered in different locations on a circular track. Multivariate analyses of the neural data indicated that information about object identity was represented secondarily to the primary information dimension of object location. In addition, the neural data related to performance on the recognition memory task. The results suggested that objects were represented as points of interest on the hippocampal cognitive map and that this map was useful in remembering encounters with particular objects in specific locations.

Device, method, and graphical user interface for manipulating user interface objects

Abstract: A computing device with a touch screen display simultaneously displays on the touch screen display a plurality of user interface objects and at least one destination object. The computing device detects a first input by a user on a destination object displayed on the touch screen display. While continuing to detect the first input by the user on the destination object, the computing device detects a second input by the user on a first user interface object displayed on the touch screen display. In response to detecting the second input by the user on the first user interface object, the computing device performs an action on the first user interface object. The action is associated with the destination object.

Method for using gesture objects for computer control

Abstract: A computer control environment introduces the Gesture environment, in which a computer user may enter or recall graphic objects on a computer display screen, and draw arrows and gesture objects to control the computer and produce desired results. The elements that make up the gesture computing environment, include a gesture input by a user that is recognized by software and interpreted to command that some action is to be performed by the computer. The gesture environment includes gesture action objects, which convey an action to some recipient object, gesture context objects which set conditions for the invocation of an action from a gesture object, and gesture programming lines that are drawn to or between the gesture action objects and gesture context objects to establish interactions therebetween.

Electronic apparatus and method for implementing user interface

Abstract: An electronic apparatus and method of implementing a user interface according to a pressure intensity of a touch on the electronic apparatus, the method including detecting a position at which the touch is input, identifying the type of object displayed on the position, and detecting the pressure intensity. Accordingly, the user can manipulate electronic apparatuses with greater convenience.

Image saliency by isocentric curvedness and color

Abstract: In this paper we propose a novel computational method to infer visual saliency in images The method is based on the idea that salient objects should have local characteristics that are different than the rest of the scene, being edges, color or shape By using a novel operator, these characteristics are combined to infer global information The obtained information is used as a weighting for the output of a segmentation algorithm so that the salient object in the scene can easily be distinguished from the background The proposed approach is fast and it does not require any learning The experimentation shows that the system can enhance interesting objects in images and it is able to correctly locate the same object annotated by humans with an F-measure of 8561% when the object size is known, and 7919% when the object size is unknown, improving the state of the art performance on a public dataset

Optimization of social distribution networks

Abstract: A computer-implemented method is provided, which includes labeling an object with one or more object terms that characterize the object. A social graph is constructed that represents an online social network, the social graph having vertices that represent respective members of the social network, and links that represent social connections between respective pairs of the members. Each of the links is tagged with one or more link terms that represent common interests between two of the members whose vertices are connected by the link. Responsively to a comparison of the object terms and the link terms with which the links have been tagged, respective distribution ranks are calculated for a plurality of the members. A message is sent to one of the members of the social network suggesting that the member distribute the object to a subset of the members responsively at least in part to the distribution ranks.

Object management method and apparatus using touchscreen

Abstract: An object management method and apparatus for a device having a touchscreen is provided for handling objects displayed on the screen with diverse multi-touch gestures. An object management method for a touchscreen-enabled device according to the present invention includes the sensing and identification of picking up at least one object displayed on the touchscreen in response to a first type multi-touch input and releasing the at least one object on the another portion of the touchscreen, or a different area or a different display on the touchscreen, in response to a second type multi-touch input. The invention includes release to the touchscreen of another device that is in wireless communication with the device having the picked up object.

Extracting data records from the web using tag path clustering

Abstract: Fully automatic methods that extract lists of objects from the Web have been studied extensively. Record extraction, the first step of this object extraction process, identifies a set of Web page segments, each of which represents an individual object (e.g., a product). State-of-the-art methods suffice for simple search, but they often fail to handle more complicated or noisy Web page structures due to a key limitation -- their greedy manner of identifying a list of records through pairwise comparison (i.e., similarity match) of consecutive segments. This paper introduces a new method for record extraction that captures a list of objects in a more robust way based on a holistic analysis of a Web page. The method focuses on how a distinct tag path appears repeatedly in the DOM tree of the Web document. Instead of comparing a pair of individual segments, it compares a pair of tag path occurrence patterns (called visual signals) to estimate how likely these two tag paths represent the same list of objects. The paper introduces a similarity measure that captures how closely the visual signals appear and interleave. Clustering of tag paths is then performed based on this similarity measure, and sets of tag paths that form the structure of data records are extracted. Experiments show that this method achieves higher accuracy than previous methods.

Similarity search of business process models

Abstract: Similarity search is a general class of problems in which a given object, called a query object, is compared against a collection of objects in order to retrieve those that most closely resemble the query object. This paper reviews recent work on an instance of this class of problems, where the objects in question are business process models. The goal is to identify process models in a repository that most closely resemble a given process model or a fragment thereof.

User interface, device and method for a physically flexible device

Abstract: A user interface, a device and a method for a physically flexible device including a flexible display (400) configured to display graphical objects (410) and a controller configured to detect a first bend and determine a resulting first foldline (411 ), determine a graphical object being intersected by the first foldline (410a, 410b) and execute a function associated with the graphical object. In another embodiment, a controller is configured to detect a bend resulting in a shape and execute a function associated with the shape. In another embodiment, a controller is configured to detect a bend of a corner and execute a function associated with the corner.

Rfid fingerprint creation and utilization

Abstract: Embodiments of the present invention provide a method, system and computer program product for creating and managing radio frequency identification (RFID) fingerprints to ensure item authenticity. In an embodiment of the invention, an object fingerprint creation method can be provided. The method can include receiving a sensed signal from at least one RFID reader resulting from at least one RFID tag affixed to a target object. The method further can include extracting a voltage produced by the RFID tag from the sensed signal. Finally, the method can include storing the voltage as a fingerprint for the target object for later comparison with another fingerprint for an authenticating object to determine whether or not the authenticating object is the target object.

Methods and system for enabling printing three-dimensional object models

Abstract: The present invention relates to a system and method of transforming an inconsistent 3D (three-dimensional) model of one or more 3D objects into a valid printable 3D model, said method comprising: (a) receiving an input model file and/or receiving modeler data, giving rise to an input 3D model having one or more parts; (b) analyzing said input 3D model and removing its one or more inconsistencies; (c) thickening the one or more parts of said 3D model, while preserving their functionality; and (d) unifying the union of interior volumes of said one or more parts, thereby removing self intersections and/or intersections between the two or more parts, giving rise to a valid printable 3D model

Image collage authoring

Abstract: A user interface that includes a catalog area, a collage mock-up area, and a mode select interface control operable to select an operational state of the user interface is displayed. Thumbnails of respective images are shown in the catalog area. A layout of a subset of the images is presented in the collage mock-up area. In response to the receipt of a user input gesture and a determination that the user interface is in a first operational state, a first action type is performed based on the type of the received user input gesture and the object type of the target object. In response to the receipt of the user input gesture and a determination that the user interface is in a second operational state, a second action type is performed based on the type of the received user input gesture and the object type of the target object.

Multi-player, multi-screens, electronic gaming platform and system

Abstract: A multi-player, multi-screens, electronic gaming platform and system is disclosed, having a game board having a main display screen for displaying a portion of the game layout that is shared by the plurality of players, and a plurality of personal game terminals, each personal game terminal assigned to a respective player and having an auxiliary display screen for displaying a portion of the game layout that is associated with the respective player. Communication between the main display and the terminals, and among terminals directly is enabled. Also enabled is seamless transfer of objects between the display screens as triggered by a game step or interactive inputs. Coded rules associated with the transferred objects can indicate where on the screen the digitized game object should appear, at what size should it be displayed, special animation or audio that should accompany the transfer of the digitized game object, etc.

Model-based and learned semantic object labeling in 3D point cloud maps of kitchen environments

Abstract: We report on our experiences regarding the acquisition of hybrid Semantic 3D Object Maps for indoor household environments, in particular kitchens, out of sensed 3D point cloud data. Our proposed approach includes a processing pipeline, including geometric mapping and learning, for processing large input datasets and for extracting relevant objects useful for a personal robotic assistant to perform complex manipulation tasks. The type of objects modeled are objects which perform utilitarian functions in the environment such as kitchen appliances, cupboards, tables, and drawers. The resulted model is accurate enough to use it in physics-based simulations, where doors of 3D containers can be opened based on their hinge position. The resulted map is represented as a hybrid concept and is comprised of both the hierarchically classified objects and triangular meshes used for collision avoidance in manipulation routines.

Searching Linked Objects with Falcons: Approach, Implementation and Evaluation

Abstract: Along with the rapid growth of the data Web, searching linked objects for information needs and for reusing become emergent for ordinary Web users and developers, respectively. To meet the challenge, we present Falcons Object Search, a keyword-based search engine for linked objects. To serve various keyword queries, for each object the system constructs a comprehensive virtual document including not only associated literals but also the textual descriptions of associated links and linked objects. The resulting objects are ranked by considering both their relevance to the query and their popularity. For each resulting object, a query-relevant structured snippet is provided to show the associated literals and linked objects matched with the query. Besides, Web-scale class-inclusion reasoning is performed to discover implicit typing information, and users could navigate class hierarchies for incremental class-based results filtering. The results of a task-based experiment show the promising features of the system.