Oscar Figueiredo

Parallelizing I/O-intensive image access and processing applications

This article presents methods and tools for building parallel applications based on commodity components: PCs, SCSI disks, Fast Ethernet, Windows NT. Chief among these tools is CAP, our computer-aided parallelization tool. CAP generates highly pipelined applications that run communication and I/O operations in parallel with processing operations. One of CAPs successes is the Visible Human Slice Server, a 3D tomographic image server that allows clients to choose and view any cross section of the human body.

Distance preserving flattening of surface sections

Curved cross-sections extracted from medical volume images are useful for analyzing nonplanar anatomic structures such as the aorta arch or the pelvis. For visualization and for performing distance measurements, extracted surface sections need to be adequately flattened. We present two different distance preserving surface flattening methods which preserve distances according to a user-specified center of interest and according to user-specified orientations. The first method flattens surface sections by preserving distances along surface curves located within planes having a user specified constant orientation. The second method flattens surfaces along curves located within radial planes crossing the center of interest. We study and compare the properties of the two flattening methods by analyzing their distortion maps. Thanks to a multiresolution approach, we provide surface flattening at interactive rates, allowing users to displace their focus point while visualizing the resulting flattened surface. These distance preserving flattening methods provide new means of inspecting curved cross-sections extracted from medical images.

Parallel unfolding and visualization of curved surfaces extracted from large three- dimensional volumes

Although many three-dimensional (3D) medical imaging visualization methods exist, 3D volume slicing remains the most commonly used technique for visualizing medical data from modali- ties such as CT, MRI, and PET. We propose to extend the possibili- ties of oblique slicing to developable curved surfaces that can be flattened and displayed in two dimensions without deformation. Such surfaces can be used to follow curved anatomical structures while preserving distance metrics at visualization time. They may also be useful for the staging of tumors, i.e., to evaluate the spatial extension of a tumor. We propose an out of core algorithm that runs in parallel on a multi-PC architecture and is able to extract surfaces from very large 3D datasets such as the visible human data set (man: 13 GB, woman: 49 GB). Experimental performance results are presented which demonstrate that parallel surface extraction is scalable and has a reasonable overhead compared with traditional oblique planar slicing. Surface extraction is made available to the public as one of the services offered by EPFLs visible human web server (http://visiblehuman.epfl.ch).

Visible Human Slice Web Server: a first assessment

The Visible Human Slice Server started offering its slicing services at the end of June 1998. From that date until the end of May, more than 280,000 slices were extracted from the Visible Man, by layman interested in anatomy, by students and by specialists. The Slice Server is based one Bi-Pentium PC and 16 disks. It is a scaled down version of a powerful parallel server comprising 5 Bi-Pentium Pro PCs and 60 disks. The parallel server program was created thanks to a computer-aided parallelization framework, which takes over the task of creating a multi-threaded pipelined parallel program from a high-level parallel program description. On the full blown architecture, the parallel program enables the extraction and resampling of up to 5 color slices per second. Extracting 5 slice/s requires to access the disks and extract subvolumes of the Visible Human at an aggregate throughput of 105 MB/s. The publicly accessible server enables to extract slices having any orientation. The slice position and orientation can either be specified for each slice separately or as a position and orientation offered by a Java applet and possible future improvements. In the very near future, the Web Slice Server will offer additional services, such as the possibility to extract ruled surfaces and to extract animations incorporating slices perpendicular to a user defined trajectory.

Digitization of Bézier Curves and Patches using Discrete Geometry

Existing algorithms for rendering BEzier curves and surfaces fall into two categories: iterative evaluation of the parametric equations (generally using forward differencing techniques) or recursive subdivision. In the latter case, all the algorithms rely on an arbitrary precision constant (tolerance) whose appropriate choice is not clear and not linked to the geometry of the image grid. In this paper we show that discrete geometry can be used to improve the subdivision algorithm so as to avoid the need for any arbitrary value. The proposed approach extends well and we present its application in the case of 2D and 3D BEzier curves as well as BEzier triangle patches and tensor-product surface patches.

New results about 3D digital lines

The current definition of 3D digital lines, which uses the 2D digital lines of closest integer points (Bresenhams lines) of two projections, has several drawbacks: the discrete topology of this 3D digital line notion is not clear; its third projection is, generally, not the closest set of points of the third euclidean projection; if we consider a family of parallel euclidean lines, we do not know how many combinatorially distinct digital structures will be built by this process; and mainly the set of voxels defined in this way is not the set of closest points of the given euclidean line. And these questions are the simplest ones; many others could be asked: dependence on the choice of the projection, intersections with digital planes, intersections between 3D digital lines,... This paper gives a new definition of 3D digital lines relying on subgroups of Z3, whose main advantage over the former one is its ability to convert any practical question into rigorous algebraic terms. It follows previously developed ideas but with a much simpler treatment and new results. In particular, we obtain a complete description of the topology of these lines and a condition for the third projection being a 2D digital line as well as a classification of digital lines of the same direction into classes of equivalent combinatorial structure.

Computer-aided parallelization of continuous media applications: the 4D beating heart slice server

Parallel servers for I/O and compute intensive continuous media applications are difficult to develop. A server application comprises many threads located in different address spaces as well as files striped over multiple disks located on different computers. The present contribution describes the construction of a continuous media server, the 4D beating heart slice server, based on a computer-aided parallelization tool (CAP) and on a library of parallel file system components enabling the combination of pipelined parallel disk access and processing operations. Thanks to CAP, the presented architecture is concisely described as a set of threads, operations located within the threads and flow of data and parameters (tokens) between operations. Continuous media applications are supported by allowing tokens to be suspended during a period of time specified by a user-defined function. Our target application, the 4D beating heart server supports the extraction of freely oriented slices from a 4D beating heart volume (one 3D volume per time sample). This server application requires both a high I/O throughput for accessing from disks the set of 4D sub-volumes (extents) intersecting the desired slices and a large amount of processing power to extract these slices and to resample them into the display grid. With a server configuration of 3 PCs and 24 disks, up to 7.3 slices can be delivered per second, i.e. 43 MB/s are continuously read from disks and 4.1 MB/s of slice parts are extracted, transfered to the client, merged, buffered and displayed. This performance is close to the maximal performance deliverable by the underlying hardware. The observed single stream server delay jitter varies between 0.6s (52% of maximal display rate) and 1.4s (92% of the maximal display rate). For the same resource utilization, the jitter is proportional to the number of streams that are accessed synchronously. The presented 4D beating heart application suggests that powerful continuous media server applications can be built on top of a set of simple PCs connected to SCSI disks.