CHP 2 Essay

CHP 2

“LITERATURE

REVIEW”

2 LITERATURE REVIEW

Image detection is conscious and physical world situation whose elements are generated by machine learning and objects input. Image detection connect the human object with computer vision. Image detection is relaible for object detection by delievering him format in form of pic, video or real time. In this object we deal with so many classes for which we have to generate the algorithm. Algorithm identify object through every possible angle. For e.g: If we have laptop for detection, our project must detect laptop in every possible angle same as human eye identification back, front, open, close, right, left.

This rule is restrict for every defined classes.

An upset in PC interface configuration is changing the way we consider PC’s. Our old generation concept that monitor or lcd is just like TV screen, must change. Other fields also dependent on machine learning just like pedestrian, mood detection, disease detection, vehicle tracking, entrance through human identification, face recognization, criminal tracking from police record.

All these highlight fields are associated with machine learning, and machine learning also relay on deep learning.

In few past years a lot of enough research is completed on machine learning, so it’s become basic issue between computer scientist and for those whose are inter-linked from it.

2.1) Introduction

Earlier in the year of 2017 Google released an API for object detection, through their TensorFlow program. This project utilizes that API to create a real-time object detection and identification program with the use of the built-in webcam from a Mac laptop. First, a model is trained on new data from a pre-trained model’s checkpoint. That model is then used in the program and makes inference on each frame that is received from the webcam. Said frame then has an overlay drawn on top and displayed to the user. The training is done on a GPU with the use of TensorFlow’s version that is specialized for GPU usage. The training is run on Windows with a GeForce GTX 980 TI.

The main perception of this project is to link human objects with computer vision. In this project user easily detect objects from his/her deliver format. This field is major part of machine learning of computer science. Not too much people are aware about this remarkable field. This field has also sub-fields. Image recoganization is also excellent fields. It contains a number of excellent research paper. This field requires enough research for new comers.

2.2) PREVIOUS WORK

Previous work in this field is not much effective, it requires more and exceptional idea in this field and project.

2.2.1) Mood Detection

Mood detection is quite previous work, this project demonstrate mood of user according to situation of home, office, street environment. Mood detection is quite effective for criminal because sometimes criminal feel something different and deliver it in investigation.

Demerit of mood detection is depend on age and gender. Angry males show higher levels of energy than angry females, It is found that males express angriness with a slow speech rate opposed to females who employ a fast speech rate

2.2.2) Medical Disease detector:

Medical detector implement especially for heart and lungs disease, these highlight disease create harmful effect on human body, if it isn’t recognize early so this would be calamitous for human’s future or present life.

Demerit of medical disease detector is when we apply machine on old age people the rays which enter in human body for detection of disaese sometimes it play role of slow poison for that person. Another way which it creates destruction is when any part of medical disease detector is not troubleshoot properly on daily basis it creates negative result and following that irrelevant report consultant give wrong medical treatment to patient

2.2.3) Vehicle Detector:

Vehicle Detector implement especially for traffic signals and car fault. In traffic signals it utilize for breaking the traffic rules and also for those violence on road and also it’s helpful to deliver those car driver whose break traffic rules.

The main demerit of vehicle detector is in bad weather which includes rainy season, darkness of clouds before raining, fog weather. Vehicle detector is also apply on that places which is free from high-rise building or commercial building or multinational organization, these building hide the camera for detection.

2.3) CURRENT WORK:

Image processing is now rapidly growing faster as compare to any other fields. In real world designer mainly connect computer vision to human daily life objects. This field is really required more and more effective work, deep learning is too much vast field in computer science. Image processing is mainly used in medical, colors changing in picture, background changing in video, dictionary learning for image set based face recognition, video survillence for animal behavior analaysis, gate recognition using active human energy Image, calculation of heart beat through BMI Image. Flowers classification using neural network based image processing.

All these highlight projects are associated with machine and deep learning. Present ear include many remarkable libarires which solve a lot of problem of designers. Tensor flow is usable library in image object detection, by importing it in our code it easily detect code with other libraries it ‘s our main library.

2.4) FUTURE ENHANCEMENT:

Future enhancement of image detection is used in various fields. Future of image object detection is apply on human fields. Better future scope of image object detection is robot, we place robot on human work place which reduce human work and human tension. If workload or tension is too much so machine can’t make mistake, as comparision from human, human always make mistake due to this mention stress.

2.5) Conclusion

These type of application used in variety of departments which increase efficiency of modern age, also keep secure humans from other attacks or hazards. These application make you modernize or efficient. These app reduce so many human effort or it also solve time taking task in few seconds or minutes. These kinds of apps keep moving forward and guiders keep playing their role in supervision form to students which motivate them toward these fields.

LIST OF PREVIOUS SOFTWARE:

S.NO APPLICATION NAME BACKEND WORKING FEATURES PROBLEMS

1 IMAGE J Generates a background image with iterations of the ranking with iterations defined by user.

Substracts the background image from the orginal image and generates a result image.

Contrast enhances the result image ImageJ can display, edit, analyze, process, save, and print 8-bit color and grayscale, 16-bit integer, and 32-bit floating point images. It can read all image file formats ImageJ supports image stacks, a series of images that share a single window, and it is multithreaded, 2 CELL PROFILER Advanced algorithms for image analysis that are able to accurately identify crowded cells and non-mammalian cell types.

A modular, flexible design allowing analysis of new assays and phenotypes. CellProfiler can read and analyze most common microscopy image formats.Biologists typically use CellProfiler to identify objects of interest (e.g. cells, colonies, C. elegans worms) it also remove high % of illumination(brightness) 3 Insight Segmentation and Registration Toolkit Generic programming, design pattern advanced C++ programming styles ITK is implemented in C++. ITK using the CMake build environment to manage the compilation process. ITK’s implementation technique of generic programming through C++ templates 4 PEDESTRIAN OPEN CV NILL OpenCV ships with a pre-trained HOG + Linear SVM model that can be used to perform pedestrian detection in both images and video streams. If you’re not familiar with the Histogram of Oriented Gradients and Linear SVM method 5 Ball Tracking / Detection with CV Convert image from RGB space to HSV space . HSV(hue saturation value) space gives us better results while doing color based segmentation. 6 AVIZO (SOFTWARE) It is basically designed for perception of 2d and 3d image processing to simulation It use neuroscience visualization. Data import, 2d/3d visualization, image processing,3d model reconstruction, Quantification & analaysis. 7 AUTO COLLAGE This software is relaible for creates a collage(technique of artvisualization), to detect face and recognize objects. 8 GIMP(general image manipulation program) It use object-oriented programming techniques. The Blend tool can be used to fill a selection with a color gradient. These color transitions can be applied to large regions or smaller custom path selections. It’s a free and open source raster graphics editor which retouch editing, free from drawing and also change image formats. 2.6) Refrences