Accelerating Pathology Image Data Cross-Comparison on CPU-GPU Hybrid Systems

Abstract

As an important application of spatial databases in pathology imaging analysis, cross-comparing the spatial boundaries of a huge amount of segmented micro-anatomic objects demands extremely data- and compute-intensive operations, requiring high throughput at an affordable cost. However, the performance of spatial database systems has not been satisfactory since their implementations of spatial operations cannot fully utilize the power of modern parallel hardware. In this paper, we provide a customized software solution that exploits GPUs and multi-core CPUs to accelerate spatial cross-comparison in a cost-effective way. Our solution consists of an efficient GPU algorithm and a pipelined system framework with task migration support. Extensive experiments with real-world data sets demonstrate the effectiveness of our solution, which improves the performance of spatial cross-comparison by over 18 times compared with a parallelized spatial database approach.

Figure 1

Cross-comparing queries for the Jaccard similarity of two polygon sets extracted from the same image.

Figure 2

Execution time decomposition of cross-comparing queries in PostGIS on a single core.

Figure 3

Polygons extracted from medical images have axis-aligned edges and integer-valued vertices.

Figure 4

The principles of PixelBox.

Figure 5

A sampling box’s position relative to a polygon: (a) outside; (b) inside; (c, d) hover.

Figure 6

A cross-comparing pipeline with dynamic task migrations.

Figure 7

Performance comparison of GEOS and PixelBox.

Figure 8

Performance of two algorithm decisions: using sampling boxes and computing areas of union indirectly.

Figure 9

Performance impact of various optimization techniques in algorithm implementation.

Figure 10

The sensitivity of pixelization threshold T.

Figure 11

Performance benefits of dynamic task migration.

Figure 12

The overall performance of SCCG compared with PostGIS-M on 18 data sets.