CHAPTER 1
INTRODUCTION
1.1. Background to the Study
As a scientific discipline, computer vision is concerned with the extraction of information from images to be employed in a decision making process. The image data can take many forms, such as video sequences, still images, digitized maps, diagrams and sketches. These images may be in colour format, grey scale or in binary format. The common approach is to extract characteristic features from the image either in the spatial domain or in some suitable transform domain. Whether the goal is classification or recognition, a measure of similarity or distance must be formulated and the success rate of the system evaluated. Some of the application areas of computer vision are:
1.1.1. Robotics
The visual recognition of image patterns is a fundamental human attribute that uses the eye as a sensor and dedicated parts of the brain as the decision making processor. The visual recognition enables humans to perform a variety of tasks such as target identification, ease of movement, handling tools, communication among others. Advances in sensing and visual perception techniques have enabled some of these attributes to be transferred to robots. For example, identification of colour is a useful asset in a robot in an industrial automation process that involves sorting. [1], [2], [3]
1.1.2. Industrial Inspection
Industrial inspection is an area in which pattern recognition is of importance. A pattern recognition system captures images
The perceptual process enables us to perceive the world through our senses of sight, smell, sound, taste, and touch. In particular, our visual system processes vast amounts of information in its environment. Rather than perceiving elements separately, our brain organizes patterns, objects, and shapes into whole forms that we can understand. The gestalt grouping principles of visual perception describe this organization as a set of principles that explain how we perceive and organize visual stimuli.
Sajda, Paul, and Leif Finkel. "Intermediate-Level Visual Representations and the Construction of Surface Perception." Web. 8 July 2015.
In her article, Kathryn Thornton, looks at the future of robotics, as of now robots depend too much on people for commands and repairs, but in the near future that can certainly change, she mainly focuses on explaining the potential that robotic science has and even explain what defines a robot. This article was published by national geographic and has been well researched by the author. This article contains pertinent information on the fairly new field of robotic engineering and what the future may hold for this science. This source shows both the positive and negative effects that robotics could potentially have on our society, the article reinforces the idea of robots being a major future innovation which could potentially save countless
In the article, “The Robot Invasion” by Charlie Gillis, the author discusses how robots have grown from the early 1990’s, to the present, and what is expected in the future. He discusses how the beginning of robotics went from simple robots to more complex robots. For example, Andrew Vardy, a computer science professor, recently created toy-sized robots that could see and was able to sort pucks by color (Gillie, 2012, p. 478). The biggest challenge faced with the growth of robots is the capability to respond to changes in the surroundings. However, there is hope for robots evolving because James McLurkin, a professor from Rice University, programmed robots smaller than a hamburger with simple combinations of tasks like following the leader, escaping deadlocks, and even cleaning floors of bread crumbs (p.479) Since 2005, a Massachusetts based company Kiva Systems, has been selling automated warehouse systems using hundreds of robots that are able to get merchandise from storage to shipping (p. 479). The warehouse system can triple distribution productivity. Humanoid robots are being used by the military field, due to their high market price. Baxter, is the first humanoid industrial use robot available for approximately $22,000.00. Many people are skeptical about the growth in technology, when it comes to robots, because they fear they could replace jobs that humans perform. Thomas Frey, an American Futurist, doesn’t look at robots taking over human jobs. He sees advancements for humans to adapt to the future of robots, as robot designers, engineers, and dispatchers. He feels humans and robots can grow together.
Recognition requires the system to search through many sets of resources stored and choose the one that corresponds to the unknown person presented. Various types of biometric systems are used for real-time identification; the most popular are based on face recognition and fingerprint recognition. However, there are other biometric systems that utilize iris and retinal scan, voice, gesture recognition, and hand geometry. Biometric technologies are becoming the basis for a wide range of solutions for personal identification and verification of high security. The basic idea behind biometrics is that our bodies contain unique properties that can be used to distinguish it from others. A biometric system is essentially a
Images with higher similarity than threshold are returned as an output. System which uses the content of the object for Information Retrieval is the CBIR.
Color is widely remarked as one of the most demonstrative visual features, and as such it has been largely studied in the context of CBIR, thus number one to a rich variety of descriptors. As traditional color features used in CBIR, there are color histogram, color correlogram, and dominant color descriptor (DCD) [1,3,4]. A simple color similarity between two images can be
Here biometrical information about individual is stored. And for gaining information a sensor is used which work as an interface between the real world and the system. When the information detected by sensor then it is compared with the information which are already available in the system. The second block performs several pre-processing as to remove old objects from the sensor for the improvement of input like removing background noise. The third block extracts significant features. The third block is an important block because the correct features need to be extracted in the best way. A template is a construction of applicable characteristics which are extracted from the source. For the creation of template an image is used with particular properties. Elements which are not used in the comparison algorithm are throw away in the template to reduce the file size.
Identification systems has received and increasing attention from security point of view. These systems rely on three main elements 1) Attribute identification 2) Biographical identification and 3) Biometric
It required a lot of study on previous work and some study of related topics which uses a little bit same technique to process their features for recognizing a person. Most of the literatures were based on palm image processing and some of the books and research papers were based on face recognition. As such, a lot of research work has been done by Chinese researchers so far. A series of researches gave me a set of methods and a set of features to be selected among them.
Thirdly, calculate the position of each object point in the images by applying a correlation algorithm with the use of a stochastic intensity pattern on the object surface.
Face is a analyzable multi-dimensional visual model and processing a process model for face recognition is challenging. This paper presents a methodological analysis for face identification based on content explanation formulation of coding and decoding the face image. categorization using the Euclidian distance. The content is to use the system for a particular face and separate from a large number of stored faces with some real time variations as well. The Eigen face attack uses particular faces with some real time variation. The Eigen face formulation uses principal components analysis (PCA) algorithm for the acceptance of the images. It gives us prompt way to insight the lower dimensional space.
Image mining systems can discover meaningful information or image patterns from a huge collection of images. Image mining determines how low level pixel representation consists of a raw image or image sequence can be handled to recognize high-level spatial objects and relationship [14]. It includes digital image processing, image understanding,
Implementation of any system needs the study of features, it may be symbolic, numerical or both. An example of a symbolic feature is color; an example of numerical feature is weight. Features may also result from applying a text extraction algorithm or operator to the input data. The related problems of feature selection and feature extraction must be addressed at the outset of any text recognition system design. The key is to choose and to extract features that are computationally feasible and reduce the problem data into a manageable amount of information without discarding valuable information.
Nowadays retail video analytics has moved out ahead of the traditional domain of loss prevention and security by providing retailers understanding business intelligence for instance queue data and store traffic statistics. Such information allows on behalf of optimized store performance, enhanced customer experience, ultimately higher reduced operational costs and profitability. This paper gives an overview of various camera-based applications in retail. It also presents a number of the promising technical guidelines for survey in retail video analytics.