Search

Capturing of infrared images is an easy task but perceptual visualization is difficult due to environmental conditions such as light rain, partly cloudy, mostly cloudy, haze, poor lightening conditions, noise generated by the sensors, geographical distance and appearances of the objects. To improve the human perception and quality of the infrared images for further processing like image analysis, image enhancement is an essential process. This paper provides a detailed review of various image enhancement techniques from contrast stretching to optimization methods used in infrared images. It also discusses the existing infrared image enhancement techniques as group such as histogram based methods, filter based methods, transform domain based methods, morphological based methods, saliency extraction methods, fuzzy based methods, learning methods, optimization methods and its popular algorithms also address the countless issues. Some of the existing image enhancement methods (Histogram Equlization, Max-median filter, Top-Hat transform) and infrared image enhancement methods (multi-scale top-hat transform, adaptive infrared image enhancement) are implemented along with the adaptive fuzzy based infrared image enhancement method and its obtained results evaluation is done on subjective and objective ways. From the results observed that the fuzzy based method works well for both subjective and objective evaluation. The paper aims to provide a complete study on image enhancement techniques and how they specially utilized while dealing with infrared images. In addition, the paper helps the researchers to select the suitable infrared image enhancement techniques for their infrared image application needs.

Eye is an essential sensory organ in human physiological system and disease in eye severely affects the vision system. Age-related-macular-degeneration (AMD) is a common eye disease in elderly people (age > 60 years) and the untreated AMD will cause severe eye illness including permanent vision loss. This research work proposes a machine-learning-system (MLS) to detect the AMD using the fundus-retinal-images. The various phases of the MLS employed in the proposed work is as follows; (i) Image thresholding based on Shannon’ Entropy and Bat Algorithm (BA + SE), (ii) Gaussian-filter (GF) based image smoothening, (iii) feature extraction, (iv) Feature selection using statistical test, and (v) classifier implementation and validation. The image features, such as the gray level co-occurrence matrix and entropies are considered to categorize the retinal image database into AMD/Non-AMD class. The performance of classifiers, such as Naive–Bayes (NB), decision-tree (DT), K-nearest neighbors, random-forest (RF) and support-vector-machine with linear kernel (SVM) are compared and the experimental outcome confirms that, the classification accuracy of SVM is superior (> 93%) compared to other classifiers.

In the proposed work, hidden Markov model (HMM) has been deployed to improve the learner's performance or grades on the basis of their psychological and environmental factors like connect/gather isolation, pleasure/comfort, depression, trust, anxiety, proper guidance, improper guidance, entertainment, and stress. The categorization of psychological and environmental factors has been done on the basis of two factors as positive and negative. The responsibility of the positive factor is to boost up learner's performance or grades, whereas negative factors reduce learning performance respectively. Finally, this paper addresses the application of HMM to determine the optimal sequence of states for different states as grades A, B, and C for different emission observations. The states identification leads to training the HMM model where optimal value of individual states computed using different observation sequences which determines the probability of state sequences. The probability of achieved optimal states is shown in different logical combinations where best state is searched among available different states using different search techniques. The computational results obtained after training are encouraging and useful.

In recent years, the Internet of Things (IoT) has become an emerging technology, which is gaining popularity among researchers and practitioners. Due to its multidisciplinary approach, IoT has been phenomenal in revolutionizing many aspects of traditional healthcare paradigms. At the same time, traditional healthcare systems can no longer satisfy the needs of a continuously growing and developing society.

Image stitching is an important task in image processing and computer vision. Image stitching is the process of combining multiple photographic images with overlapping fields of view to produce a segmented panorama, resolution image. It is widely used in object reconstruction, panoramic creating. In this paper, we present an approach based on deep learning for image stitching, which are applied to generate high resolution panoramic image supporting for virtual tour interaction. Different from most existing image matching methods, the proposed method extracts image features using deep learning approach. Our approach directly estimates locations of features between pairwise constraint of images by maximizing an image- patch similarity metric between images. A large dataset high resolution images and videos from natural tourism scenes were collected for training and evaluation. Experimental results illustrated that the deep feature approach outperforms.

Fuzzy logic is an arrangement to sort with registering in light for “degrees of truth” instead of the standard thing “genuine or false” (1 or 0) Boolean logic. A Deoxyribonucleic corrosive (DNA) microarray (in like manner normally known as Deoxyribonucleic Acid chip or bio-chip) is a collection of small DNA-spots annexed to a hard surface (Mukhopadhyay et al., Analysis of microarray data using multiobjective variable string length genetic fuzzy clustering, 2009) [1]. DNA microarrays are used to gage the explanation hierarchy of significant amounts of characteristics in the meantime or to geno-type distinctive areas of a genome. Each Deoxyribonucleic Acid spot consist of micro-moles (appx. 11 mol) of a particular Deoxyribonucleic Acid progression, justified as tests (or writers or oligos). It can be a small territory of a quality or other DNA segment which is used to hybridize a complementary DNA/complementary RNA (in like manner called antagonistic to identify Ribonucleic Acid) test (called centre) under high-stringency terms. Test object hybridization is normally recognized and acquired through distinguishing proof of “fluorophore-, silver-, or chemiluminescence”—named centres to choose relatively abundant nucleic destructive game plans for the goal. The principal nucleic destructive displays were expansive scale groups around 9 cm × 12 cm and the fundamental modernized picture based examination was circulated in 1981. Microarray information investigation includes a few unmistakable advances. Changing any of the means will change the result of the examination, so the MAQC Project was made to recognize an arrangement of standard procedures (Mukhopadhyay et al., Analysis of microarray data using multiobjective variable string length genetic fuzzy clustering, 2009) [1], (Li, International Seminar on Future BioMedical Information Engineering, 2008) [4]. Organizations exist that utilization the MicroArray/Sequencing Quality Control (MAQC) conventions to play out a total examination. The human genome contains roughly 21,000 qualities. At any given minute, each of our cells has some mix of these qualities turned on, and others are killed (Pujari, Data Mining Techniques, 2001) [2].

The contrast enhancement of mammograms at preprocessing stage optimizes the overall performance of a computer-aided detection (CAD) system for breast cancer. In the proposed approach, contrast enhancement is performed using a sigmoidal transformation mechanism followed by extracting a set of 14 Haralick features. For classification purposes, a support vector machine (SVM) classifier is used which sorts the input mammogram into either normal or abnormal subclasses. The performance of the classifier is estimated by calculating parameters like accuracy, specificity, and sensitivity. The performance of the proposed approach has been reported to be better in comparison to other existing approaches.