Article Preview
Top1. Introduction
In digital image processing we process the images according to our need. Image enhancement means to improve the visual quality of an image which depends on the application circumstances. In this paper, an electron microscopic image enhancement for feature extraction using block based adaptive contrast enhancement using modified sigmoid function algorithm is proposed. Adaptive image enhancement and feature extraction techniques are widely used in applications, where subjective quality of an image is important perspective for colony of micro-organism. In recent days, adaptive image enhancement has emerged as effective tool in the field of medical science and microbiology to extract the features from an image. Particularly in microbiology field, the actinomycetes images need enhancement adaptively for feature extraction.
For image enhancement R. Gonzalez et al. (1992), J. Stark (2000) and V. Caselles et al. (1998) suggested an automatic technique which is known by Histogram Equalization (HE). J. Russ (2002) found that HE technique is not working effectively and efficiently when contrast statistics changes across the image. R. Gonzalez et al. (1992), V. Caselles et al. (1998), Pizer et al. (1987) and Alex Stark et al. (2000) proposed an Adaptive Histogram Equalization (AHE) technique, which removes this drawback by generating the mapping for each pixel from the histogram in a neighboring window. Alex Stark et al. (2000) and Zuiderveld K (1994) suggested that the AHE has limitation that does not allow the degree of contrast enhancement to be regulated. Alex Stark et al. (2000) discussed the Global Histogram Equalization (GHE), which possesses the problem of excessive enhancement and V. Caselles et al. (1998) and Kim et al. (2008) the Local Histogram Equalization (LHE) causes blocking artifacts. Nowadays several techniques are available which are based on histogram equalization proposed by various researchers. These methods may lead to over enhancement or other artifacts like blocking, flickering and contouring. Global histogram equalization adjusts the intensity histogram to approximate a uniform distribution of pixels in each gray level. It equally works on all regions of an image and thus often provides poor local performance in terms of information preservation of an image. And because of this, some important features of actinomycetes images are lost. To overcome these problems, an adaptive image enhancement and feature extraction algorithm has been proposed in this paper, which is based on block of modified sigmoid function. Goodfellow et al. (1984) and Embley et al. (1994) suggested that actinomycetes are microscopic, aerobic (bacteria grow in the presence of oxygen), anaerobic (grow in the absence of oxygen), Gram positive, filamentous bacteria which contain high G-C content about 69-78%. Hoit (1994) and Ensign (1978) distinguished it from other bacteria by their branching vegetative hyphae or aerial mycelium produced chains of arthospores. Cross et al. (1982) suggested the actinomycetes are well known for their ability to produce antibiotics, metabolities, aminoacids, enzymes and other bioactive compounds. From the prespective of Crawford et al. (1988), they are also known for their ability to degrade organic and inorganic complexes, cellulose and lignin. Knorreet al. (2000) suggested that it utilise an inorganic nitrogen molecules. Li et al. (1995) told that the actinobacteria occur widely in various habitat, including manmade environments. Li et al (2018) proposed the fourth order moment for contrast enhancement. Shyam Lal et al. (2014) proposed the ACEBSF technique, which provides Measure of Enhancement Factor (EMF) values for image1, image2, image3 and image4 are 1.58, 1.85, 1.53 and 1.58 respectively. Our proposed technique provides superior results as compared to ACEBSF technique of Shyam Lal et al. (2014) and other existing techniques. The value of EMF is enhanced by reducing the size of block of modified sigmoid function. The proposed algorithm avoids the excessive enhancement and makes the contrast enhancement adjustable and maintain the brightness level adaptively. This paper is organized as follows: Section 1 provides literature review and Section 2 provides a brief discussion of sigmoid function. Section 3 describes proposed algorithm for contrast enhancement and feature extraction of electron microscopic images. Section 4 gives simulation results and discussions to demonstrate the performance of proposed algorithm for microscopic images. Finally, the conclusions are drawn in section 5.