Optimized execution of morphological reconstruction in large medical images on embedded devices
- PDF / 1,248,026 Bytes
- 13 Pages / 595.276 x 790.866 pts Page_size
- 23 Downloads / 136 Views
ORIGINAL RESEARCH PAPER
Optimized execution of morphological reconstruction in large medical images on embedded devices Felipe Cabral1 · Oscar Anacona‑Mosquera2 · Renato C. Sampaio2 · George Teodoro3 · Carlos H. Llanos2 · Ricardo P. Jacobi2 Received: 7 April 2020 / Accepted: 20 August 2020 © Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract This work presents a hardware/software co-design implementation of the morphological reconstruction targeting a Systemon-Chip (SoC) FPGA-based embedded system. Our approach processes large images with fast algorithms. This was achieved by the proposal and use of an execution scheme that partitions the input image into sub-images that are independently processed before a second phase is executed to enable propagation of information among sub-images. The SoC is efficiently used by processing sub-images on hardware (the costly phase), while the software takes care of computations due to discontinuities that are irregular and inefficient for the hardware execution. Several optimizations were proposed, including parallel software and hardware execution and the use of borders to minimize computations in the discontinuities correction. This enables the processing of large images from our use-case brain cancer tissue image analysis application. For an image of 8192 × 8192 pixels, our co-design solution attains a speedup of 12.7 × vs. the software execution (Dual core ARM A9 Cortex). Keywords Morphological reconstruction · FPGA · System-on-Programmable-Chip · Image processing
1 Introduction Mathematical morphology is a theory employed in analysis and processing of spatial structures in images [1]. Among the several mathematical morphology operations available, we are interested in the morphological reconstruction. It * Ricardo P. Jacobi [email protected] Felipe Cabral [email protected] Oscar Anacona‑Mosquera [email protected] Renato C. Sampaio [email protected] George Teodoro [email protected] Carlos H. Llanos [email protected] 1
Qualcomm, Cork, Ireland
2
Universidade de Brasília, Brasilia, Brazil
3
Universidade Federal de Minas Gerais, Belo Horizonte, Brazil
belongs to a set of geodesic transformations [1], which is frequently employed in shape information extraction in images from several domains. We targeted morphological reconstruction, because it is the most costly operation in our motivating pathology image analysis application [2, 3]. This application analyses highresolution tissue images from several cancer types in which it segments nuclei in images and correlates their structures with disease sub-type, survival, etc. The acceleration of morphological reconstruction in FPGAs is a major step to enable efficient processing of those tissue images in portable devices. This would allow for an expert pathologist to quickly visualize results that could assist, for instance, in selecting precision treatments [2, 3]. The real-time processing requirement here is in the order of seconds, which, for instance, is sufficient for a pathologist interacti
Data Loading...
