Datasets
Standard Dataset
Noise based in speckle
- Citation Author(s):
- Submitted by:
- miguel vera
- Last updated:
- Fri, 11/17/2023 - 17:26
- DOI:
- 10.21227/9tr1-q453
- License:
102 Views- Categories:
- Keywords:
Abstract
This synthetic dataset or phantom consists of 3 jpg format databases, in the two-dimensional (2-D) domain, which are identified as follows:
DB1: Ground Truth
DB2: Speckle noise with zero mean and 0.005 standard deviation
DB3: Speckle noise with zero mean and 0.05 standard deviation
In order to generate all synthetic DBs, specific gray levels were considered, and the myocardium and LV wall was simulated by an inner circle and an outer circle, respectively. The spatial resolution of each image was 256 × 256 pixels. TAll the databases were built in a set of concentric circular circles or layers, which were conditioned to contain densitometric information obtained by considering the average gray levels (GL) exhibited by a real cardiac ultrasound database. To introduce in the databases the information related to the GLs, each of the generated circular layers was processed with a filling technique based on a region growing algorithm.
The dataset considers two concentric circles in order to simulate the human heart endocardium and the wall of the main heart chamber called the left ventricle.
The 2-D computer-generated cardiac phantom is shown as a double circle array.
Each database consists of 50 images with intensity information (gray levels), is quantized to 8 bits and its spatial resolution is 256 × 256.
