Image processing-based fuzzy inference system for Penicillium and Aspergillus species imaging

Fungi is a massively diverse organism that is widely distributed in the ecosystem. Due to its widespread distribution, fungi intervenes with the human life as it can be beneficial in the medicinal field in producing drugs. However, it is also harmful as it can cause a wide range of disease in humans. Due to this reason, it is important to determine whether a species of fungi can be harmful or not. Thus, an identification of fungi is crucial. The main problem in detecting a species of fungi is that a microscopic image of fungi is usually low in contrast. Plus, the structure of fungi is complex and changes according to its level of maturity. In addition, the current conventional method focuses on identifying the structure of fungi directly from the microscope instead of examining it in the form of an image. Thus, in this research, the aim is on enhancing the fungi image quality in order to see its morphological features more clearly. A new approach which combines the image enhancement techniques of adaptive histogram equalization (AHE) and adaptive gamma correction (AGC) together with the technique of fuzzy inference system is proposed. The main focus of this technique is on improving the features of brightness and contrast of a microscopic image by applying two fixed gamma values for all the fungi images for the gamma transformation function. Subsequently, an improved version of the proposed image enhancement technique is introduced by implementing the technique of fuzzy partition in place of fuzzy inference system. This new technique utilizes the concept of fuzzy partition together with the surrounding neighbourhood technique in order to obtain a specific gamma value based on the properties of a fungi image. The proposed image enhancement techniques are then investigated to determine whether they are suitable to be implemented on the microscopic image of fungi. The performance is evaluated on the collected database consisting of microscopic image of fungi in terms of subjective and objective evaluation by visual inspection and Image Quality Assessment (IQA) metrics. The results obtained conclude that the proposed image enhancement technique, fuzzy purity gamma adaptive histogram equalization (FptGAHE) had a better performance in terms of enhancing the microscopic image of fungi in comparison with the other techniques. This is by obtaining a high percentage value of 90.64%, 97.22% and 98.04% for the performance evaluation of accuracy, sensitivity and specificity respectively.