Page 188 - Handbook of Deep Learning in Biomedical Engineering Techniques and Applications

Chapter 6 Plant leaf disease classification based on feature selection 177

• Textural features:
Textural features are the mean of gray-level cooccurrence
matrices (GLCMs)
There are 36 features in this case (Fig. 6.20)

4.3.2 Feature selection

In this study, we implement a wrapper-based method with a
hybrid metaheuristic called adaptive particleegray wolf optimi-
zation (APGWO), which is combined from particle swarm optimi-
zation (PSO) and gray wolf optimization (GWO).

4.3.2.1 Particle swarm optimization
PSO was proposed by Kennedy and Eberhart in 1995 [18] and
inspired from the social behavior of birds flocking. The algorithm
of PSO involves in multiple particles that find the best position of
particle and their own best positions. It can be classified as a
swarm intelligence algorithm such as ant colony algorithm. In
PSO, each individual searches space with a velocity, which is
dynamically adjusted according to its own flying experience and

Figure 6.20 Feature extraction: (A) original image, (B) CLAHE is applied on H
channel, (C) extracted defective regions, (D) final result.