Object (computer science)

About: Object (computer science) is a research topic. Over the lifetime, 106024 publications have been published within this topic receiving 1360115 citations. The topic is also known as: obj & Rq.

Code symbol reading system having adjustable object detection

Abstract: A code symbol reading system is provided. The code symbol reading system has an object detection subsystem that can be configured to detect only those objects that are positioned at a specified distance from the system. The object-detection subsystem thereby initiates the scanning process only when an object is detected at substantially the specified distance from the system.

ConceptBase—a deductive object base for meta data management

Abstract: Deductive object bases attempt to combine the advantages of deductive relational databases with those of object-oriented databases. We review modeling and implementation issues encountered during the development of ConceptBase, a prototype deductive object manager supporting the Telos object model. Significant features include: 1) The symmetric treatment of object-oriented, logic-oriented and graph-oriented perspectives, 2) an infinite metaclass hierarchy as a prerequisite for extensibility and schema evolution, 3) a simple yet powerful formal semantics used as the basis for implementation, 4) a client-server architecture supporting collaborative work in a wide-area setting. Several application experiences demonstrate the value of the approach especially in the field of meta data management.

Predicting human gaze beyond pixels

Abstract: A large body of previous models to predict where people look in natural scenes focused on pixel-level image attributes. To bridge the semantic gap between the predictive power of computational saliency models and human behavior, we propose a new saliency architecture that incorporates information at three layers: pixel-level image attributes, object-level attributes, and semantic-level attributes. Object- and semantic-level information is frequently ignored, or only a few sample object categories are discussed where scaling to a large number of object categories is not feasible nor neurally plausible. To address this problem, this work constructs a principled vocabulary of basic attributes to describe object- and semantic-level information thus not restricting to a limited number of object categories. We build a new dataset of 700 images with eye-tracking data of 15 viewers and annotation data of 5,551 segmented objects with fine contours and 12 semantic attributes (publicly available with the paper). Experimental results demonstrate the importance of the object- and semantic-level information in the prediction of visual attention.

Interactive Occlusion and Collision of Real and Virtual Objects in Augmented Reality

Abstract: We present several techniques for interactively performing occlusion and collision detection between static real objects and dynamic virtual objects in augmented reality. Computer vision algorithms are used to acquire data that model aspects of the real world. Either geometric models may be registered to real objects, or a depth map of the real scene may be extracted with computer vision algorithms. The computer vision-derived data are mapped into algorithms that exploit the power of graphics workstations, in order to interactively produce new effects in augmented reality. By combining live video from a calibrated camera with real-time renderings of the real-world data from graphics hardware, dynamic virtual objects occlude and are occluded by static real objects. As a virtual object is interactively manipulated collisions with real objects are detected, and the motion of the virtual object is constrained. Simulated gravity may then be produced by automatically moving the virtual object in the direction of a gravity vector until it encounters a collision with a real object.

Learning Video Object Segmentation with Visual Memory

Abstract: This paper addresses the task of segmenting moving objects in unconstrained videos. We introduce a novel two-stream neural network with an explicit memory module to achieve this. The two streams of the network encode spatial and temporal features in a video sequence respectively, while the memory module captures the evolution of objects over time. The module to build a “visual memory” in video, i.e., a joint representation of all the video frames, is realized with a convolutional recurrent unit learned from a small number of training video sequences. Given a video frame as input, our approach assigns each pixel an object or background label based on the learned spatio-temporal features as well as the "visual memory" specific to the video, acquired automatically without any manually-annotated frames. The visual memory is implemented with convolutional gated recurrent units, which allows to propagate spatial information over time. We evaluate our method extensively on two benchmarks, DAVIS and Freiburg-Berkeley motion segmentation datasets, and show state-of-the-art results. For example, our approach outperforms the top method on the DAVIS dataset by nearly 6%. We also provide an extensive ablative analysis to investigate the influence of each component in the proposed framework.