Anup Kale

Digital multimedia archiving based on optimization steganography system

As soon as digital artifacts have become a part and parcel of everyday life, the need for digital media archives with the capacity of preserving the given metadata has risen impressively. The process of converting the digital metadata to archives, however, is fraught with a number of difficulties, the key one concerning the methodology for embedding high payload capacity information into the digital multimedia and at the same time retains high quality of the image. The given paper will consider steganography as a possible solution to the aforementioned issue. Allowing for detecting the genetic algorithm for boosting the PSNR value with the information of high capacity will help solve the issue regarding the digital multimedia archiving. Many sizes of data are embeded inside the images and the PSNR (Peak signal-to-noise ratio) is also taken for each of the images verified.

Intelligent health care — A Motion Analysis system for health practitioners

In the proposed work we present a combination of two paradigms: Wireless Sensor Networks (WSN) and Computer Vision applied for Motion Analysis. In this work the Computer Vision provides high-level behavioural monitoring and analysis, whereas Wireless Sensors capture detailed parameters of a moving object. Fusion of sensory information received from both types of sensors provides micro-level and macro-level details. These combined details can be used in various application areas. In considered applications, one of the areas can be Robotics. In this case this strategy can be used to monitor health of robots under certain actions and situations. Another important application domain is health care and rehabilitation of injured persons. In this application, movement of an injured body portion is measured after its treatment. Apart from the analysis of motion we also propose optimized movement advice to patients. Optimum motion advice is very useful in case of sports injury to recover strength and performance. In this paper we produce experimental work performed by simulating different movements of hands and legs in free space. The experimental simulation provides a broad range of data on motion analysis with visualization. The third area of application that is explored is elderly patient condition monitoring and motion analysis for health monitoring.

Parallel robot vision using genetic algorithm and object centroid

Parallel Robots are playing a very important role in the medical, automotive, food and many manufacturing applications. Due to its high speed and efficient operation, it is gaining an increasing popularity in these application domains. For making the parallel robots more automated and an intelligent a machine vision system with robust performance is needed. Here, a Machine Vision Algorithm based on Genetic Evolutionary principles for object detection in the Delta Parallel Robot based systems is proposed. The solution applies a simple, robust and high speed algorithm to accurately detect objects for the application domain. The Image Acquisition of a robots workspace is performed by using a camera mounted on the end-effector of the robot. The system is trained with the object database and with the most significant visual features of every class of objects. Images are assessed periodically for detecting the Region of Interest (ROI) within an image of the robots workspace. The ROI is defined as an area in which a presence of object features is detected. The ROI detection is achieved by applying a random sampling of pixels and an assessment of color threshold of every pixel. The color intensity is assumed as one of the features for classification that is based on the training data. After classification process, the Genetic Algorithm is applied to locate the centroid of an object in every class. In a given application class, the Centroid is considered as the most important feature. Knowledge of an approximate location of the Centroid of objects helps to maintain a high speed and reliable pick and place operations of the Delta robot system. The proposed algorithm is tested by detecting presence of electronic components in the workspace. Experimental results show that the suggested approach offers a reliable solution for the Delta robot system.

Cooperative agent-based SANET architecture for personalised healthcare monitoring

The application of an software agent-based computational technique that implements Extended Kohonen Maps (EKMs) for the management of Sensor-Actuator networks (SANETs) in health-care facilities. The agent-based model incorporates the BDI (Belief-Desire-Intention) Agent paradigms by Georgeff et al. EKMs perform the quantitative analysis of an algorithmic artificial neural network process by using an indirect-mapping EKM to self-organize. Current results show a combinatorial approach to optimization with EKMs provides an improvement in event trajectory estimation compared to standalone cooperative EKM processes to allow responsive event detection for patient monitoring scenarios. This will allow healthcare professionals to focus less on administrative tasks, and more on improving patient needs, particularly with people who are in need for dedicated care and round-the-clock monitoring.

Evolutionary Feature Optimization and Classification for Monitoring Floating Objects

Water surfaces are polluted due to various man-made and natural pollutants. In urban areas, natural water sources including rivers, lakes and creeks are the biggest collectors of such contaminants. Monitoring of water sources can help to investigate many of details relating to the types of litter and their origin. Usually two principle methods are applied for this type of applications, which include either a use of in-situ sensors or monitoring by computer vision methods. Sensory approach can detect detailed properties of a water including salinity and chemical composition. Whereas, a camera based detection helps to monitor visible substances like floating or immersed objects in a transparent water. Current computer vision systems require an application specific computational models to address a variability introduced due to the environmental fluctuations. Hence, a computer vision algorithm is proposed to detect and classify floating objects in various environmental irregularities. This method uses an evolutionary algorithmic principles to learn inconsistencies in the patterns by using a historical data of river pollution. A proof of the concept is built and validated using a real life data of pollutants. The experimental results clearly indicate the advantages of proposed scheme over the other benchmark methods used for addressing the similar problem.

HyMuDS: A Hybrid Multimodal Data Acquisition System

This paper outlines an architectural perspective for a multimodal data acquisition to be implemented in order to monitor contamination in urban waterways. The purpose is to develop an approach to detect objects and anomalies in dynamic environmental conditions. For overcoming effects of environmental variations like high reflectivity, heat waves, fog and variable illumination, an implementation with multiple camera modalities including infrared, ultraviolet and visual spectrum is proposed. Detection of a micro-level parameters related with the environment and the water, analog sensing nodes connected to a wireless gateway are deployed. Main parameters under consideration include temperature, salinity, moisture and illumination. Software architecture for a data acquisition purpose is implemented in a C# .Net development environment. This software implementation allows parallel or concurrent data acquisition operate with a greater efficiency. Another important aspect of the software architecture implemented is to allow use of heterogeneous data for post-acquisition analysis. A problem specific data storage approach is proposed and implemented to improve availability and accessibility of the acquired data. Initial trials of this system clearly indicate merits of the system proposed. This approach has a strong capability to support capturing huge information of different scenarios and with a greater efficiency.

Managing Dynamism of Multimodal Detection in Machine Vision Using Selection of Phenotypes

Multimodal Sensor Vision is a technique for detecting objects in dynamic and uncertain environmental conditions. In this research, a new approach for automated feature subset selection-mechanism is proposed that combines a set of features acquired from multiple sensors. Based on changing environmental conditions, the merits of respective sensory data can be assessed and the feature subset optimized, using genetic operators. Genetic Algorithms (GAs) with problem specific modifications improve reliability and adaptability of the detection process. In the new approach, a traditional GA is customized by combining the problem profiled encoding with a specialized operator. Application of an additional operator prioritizes and switches within the feature subsets of the algorithm, allowing a feature level aggregation that uses the most prominent features. The approach offers a more robust and a better performing Machine Vision processing.

iMuDS: An Internet of Multimodal Data Acquisition and Analysis Systems for Monitoring Urban Waterways

Freshwater monitoring is becoming an essential activity due to limited availability of drinking water and an increasing presence of various pollutants. Tons of toxic waste added to water sources everyday contributes to the decrease in the planet’s biodiversity and even an extinction of many species of animals and marine life. Many millions of birds perish each year due to waterway pollution. New technologies such as the Internet of Things (IoT), Wireless Sensor Networks and computer vision allow us to monitor fresh water sources in a continuous mode. To minimize the effects of pollution, various monitoring activities can be planned and executed for very large areas and geographical regions. This work presents a system architecture for the IoT-based multimodal data acquisition and analysis system. The idea is to deploy sensor clusters in various locations of a waterway to create a network of sensing and measuring smart devices. Every cluster of such devices can be perceived as a ‘thing’. Such a ‘thing’ or a node has camera sensing modalities for a macro level pollution detection with analog sensors to measure microlevel water parameters. Our solution involves a low power microprocessor devices provisioned to capture raw data, extract features from the raw data and then transmit these data to the Cloud for further analysis and reporting. A 5G mobile network communication is used for data transmission. The Cloud server runs a software framework that supports a sophisticated analysis and trending of various environmental parameters such as surface density of water, salinity, temperature, etc. The proposed software framework has a set of computational algorithms to process features supplied by each node. These algorithms can classify features into various classes like floating objects, water salinity level, etc. An experiment to simulate the ‘IoT’ data acquisition is conducted to validate the proposed solution. Based on a case study, this solution can be used in a real-life scenario representing as a feasible and viable solution to track pollution in urban waterways.