Classification gel electrophoretic image of DNA Fusarium Graminearum featuring support vector machine

Fusarium Graminearum is best known as plant pathongen and most commonly found on cereal grains, wheat and barley. It has the detrimental interactions with various grains, causing numerous diseases such as gibberella ear rot and head blight. This study is to detect the presence of F. Graminearum in plant via image processing and artificial intelligence. The standard DNA gel electrophoresis procedures are used in image formation while machine learning is achieved by means of homomorphic filtering and support vector machine (SVM). Meanwhile the Gray Level Co-occurrence Matrix (GLCM) is used in feature extraction. On average, the methods and procedures returned a correct classification rate of more than 97%, with both sensitivity and specificity of 97.5%. This study paves the way for development of an imaging system to detect other types of pathogenic microbes in plants and food materials electronically.