Pixel Classification Based on Local Gray Level Rectangle Window Sampling for Amniotic Fluid Segmentation

Abstract

This study analyses the use of a pixel classification model to segment amniotic fluid areas on ultrasound (US) images characterized by noise, blurry edge, artifacts, and low contrast. In contrast with the previous methods, this study constrains a training set of pixels based on neighbourhood information with the rectangle window sampling method used to determine the characteristics of each pixel in its environment specifically. The feature extraction is no longer based on the global characteristics of the object rather by taking the value of each pixel in the object area using the sampling window. This research also combines the local first-order statistical methods and gray level information in the window to obtain the characteristics of each pixel. Furthermore, Random Forest and Decision Tree (C.45) were used to classify each pixel into four classes, namely amniotic fluid, placenta, uterus, and fetal body. The classification performance testing of pixel sampling data showed that the Random forest with 5 × 7 window sizes achieved the highest performance at 99.5% accuracy, precision, and recall, respectively. Furthermore, the proposed model was evaluated using 50 new test US images to segment the amniotic fluid area. According to experimental result, proposed models can produce better segmentation area with an increase in the IoU value by 18.3% or a Jaccard coefficient value rate of 0.183 in the range of 0-1 with the previous state of the art method. Furthermore, the proposed model reduces the error rate and improves accuracy by 6.61% and 84.77%, respectively.

Combining CNN Feature Extractors and Oversampling Safe Level SMOTE to Enhance Amniotic Fluid Ultrasound Image Classification

“…Meanwhile, parameters for measuring the performance of AFV classification employed accuracy, precision, and recall, as indicated in Eqs. ( 15)-( 17) [12].…”

“…Another study [11] implemented a deep learning (DL) network comprising AF-Net and an auxiliary network, which used a dataset of 2380 ultrasound images, achieving a DSC performance of 0.8559. Furthermore, [12] performed segmentation on 50 datasets using the proposed pixel classification method and achieved a DSC performance of 0.814 by comparing local window techniques. A previous study [13] also proposed pixel classification with a specified window size limit and in combination with several feature extractions such as gray-level, gray-level-local variance, and distance angle pixel to identify the amniotic fluid area, yielding a DSC performance of 0.876.…”

Modified of Single Deepest Vertical Detection (SDVD) Algorithm for Amniotic Fluid Volume Classification

“…The experiment was conducted on 40 amniotic fluid testing images and the unit of length was centimeters (cm). Meanwhile, the parameters used to evaluate the performance of amniotic fluid volume classification included accuracy, precision, and recall, as indicated in equations (1-3) [20].…”

Amniotic Fluid Classification and Artificial Intelligence: Challenges and Opportunities

“…The results obtained by the model had a good accuracy of 0.9052, outperforming many studies. In another study, Ayu and Hartati [13] investigated a pixel classification algorithm to differentiate AF regions on US pictures with noise, hazy edges, distortions, and poor contrast. The method involved local first-order statistical methods and data as gray-level to produce each pixel's traits.…”