Large Object Space Research Articles

Progress Overview of Capturing Method for Integral 3-D Imaging Displays

An integral 3-D technique provides a 3-D spatial image viewable from varying positions without the use of special light sources or viewing glasses. Therefore, this technique shows promise for diverse applications in various fields, including 3-D television broadcasting, advertising, and medical diagnostics. However, there are problems in capturing and displaying large amounts of information in realizing practical integral imaging devices. This paper overviews integral 3-D capturing methods and analyzes integral 3-D imaging technology at its capturing and displaying stages. To overcome the resolution problem, it also introduces our recent work for capturing high-resolution integral imaging information. The introduced device consists of a multiple-lens array and a complementary metal–oxide–semiconductor image sensor with a circuit patterned using multiple exposures. This device can capture depth-controlled spatial information by introducing additional optics. Two types of optics for depth control are applied to the capturing device: one functions as a convex lens to control and compress a relatively large object space and the other functions as an afocal lens array that controls a relatively small object space without any distortion in the depthwise direction. Experimental results of spatial information capturing and 3-D image displays confirm that the method produces 3-D images having an appropriate motion parallax. The presented method is scalable; thus, this technique offers possibilities for developing advanced high-resolution integral 3-D imaging devices.

Packer: Parallel Garbage Collection Based on Virtual Spaces

The fundamental challenge of garbage collector (GC) design is to maximize the recycled space with minimal time overhead. For efficient memory management, in many GC designs the heap is divided into large object space (LOS) and normal object space (non-LOS). When either space is full, garbage collection is triggered even though the other space may still have plenty of room, thus leading to inefficient space utilization. Also, space partitioning in existing GC designs implies different GC algorithms for different spaces. This not only prolongs the pause time of garbage collection, but also makes collection inefficient on multiple spaces. To address these problems, we propose Packer, a parallel garbage collection algorithm based on the novel concept of virtual spaces. Instead of physically dividing the heap into multiple spaces, Packer manages multiple virtual spaces in one physical space. With multiple virtual spaces, Packer offers efficient memory management. With one physical space, Packer avoids the problem of an inefficient space utilization. To reduce the garbage collection pause time, we also propose a novel parallelization method that is applicable to multiple virtual spaces. Specifically, we reduce the compacting GC parallelization problem into a discreted acyclic graph (DAG) traversal parallelization problem, and apply it to both normal and large object compaction.

A study of large object spaces

This paper examines the design space for copying garbage collectors (GCs) in which "large objects" are managed in a separate, non-copy-collected space. We focus on two main issues:1. how to set the policy for classifying objects as "large"2. how to manage the large object spaceWe explore these issues experimentally using the Oscar GC testbed. In particular, we vary the threshold size of large objects and also whether the objects may contain pointers. Furthermore, we compare the performance of treadmill collection to that of mark-and-sweep collection for managing the large object space.We find that for some heaps there is a minimum cutoff size below which adding objects to the large object space does not result in a performance improvement, while for others no such cutoff exists. In general, including pointer-containing objects in the large object space seems beneficial. Finally, the exact method used to collect the large object space does not significantly influence overall performance.

A Video Technique for Obtaining 3-D Coordinates in Alpine Skiing

A video technique to obtain 3-D data in an Alpine skiing competition was investigated. The flight and landing phases of a jump were recorded during the 1994 Olympic combined downhill race. A direct linear transformation (DLT) implementation was applied, which computes the DLT parameters for each video image of each camera separately. As a consequence, one is able to pan and tilt the cameras and zoom the lenses. The problem of distributing control points in the large object space could be solved satisfactorily. The method proved to be suitable for obtaining 3-D data with reasonable accuracy, which is even sufficient for inverse dynamics. The computed resultant knee joint forces and moments compare well with results reported by other authors.

THRESHOLD VOTING IS FUNDAMENTALLY SIMPLER THAN PLURALITY VOTING

We show that n-way plurality voting on a large unordered object space has time and space complexities of θ(n2) and θ(n), respectively. If the object space is ordered, then sorting can be used to reduce the time complexity to the optimal θ(n log n). We then prove that weighted t- out-of -∑vi threshold voting on such an object space has time complexity O(np) and needs working storage space for only p objects, where p=[(∑vi)/t]. Thus, unless t is very small, threshold voting is considerably simpler than plurality voting. As a corollary, weighted majority voting can be performed in linear time with working storage for a single input object.

A third generation Smalltalk-80 implementation

A new, high performance Smalltalk-80™ implementation is described which builds directly upon two previous implementation efforts. This implementation supports a large object space while retaining compatibility with previous Smalltalk-80™ images. The implementation utilizes a interpreter which incorporates a generation based garbage collector and which does not have an object table. This paper describes the design decisions which lead to this implementation and reports preliminary performance results.