Open AccessResearch on Application of Improved Random Forest in Medical Ultrasound Image Classification
Author(s) -
Min Cui,
Zhu Haijiang
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1584/1/012007
Subject(s) - pixel , rheumatoid arthritis , ultrasound , random forest , artificial intelligence , medicine , synovitis , pattern recognition (psychology) , grayscale , computer science , computer vision , radiology
Gray-scale ultrasound imaging methods are commonly used to assess synovitis in Rheumatoid Arthritis (RA) in clinical practice. This paper proposed an improved classification algorithm for ultrasound image of metacarpophalangeal joint in Rheumatoid Arthritis(RA). Using several features of gray scale co-occurrence matrix and improved Random Forest method to achieve automatic classification of Rheumatoid Arthritis(RA). Three grading experiments were carried out in the experiment: the first one is for the binary classification of grade 0 (normal) and grade 3 (lesion) ultrasound images of 80 × 50 pixels, 130 × 70 pixels, and 185 × 90 pixels. The second is for the binary classification of grade 0 (normal) and grade 1/2/3 (lesion) ultrasound images of 80 × 50 pixels, 130 × 70 pixels, and 185 × 90 pixels. The third is for four-grade classification of ultrasound images at grade 0(normal), grade 1(mild), grade 2(moderate), and grade 3(severe). Conclusion shows that use the method of combing gray scale co-occurrence matrix and improved Random Forest algorithm can achieve automatic classification of Rheumatoid Arthritis(RA), which has a high Classification Accuracy.