Computer Image Processing: Estimating Performance and Algorithms

This presentation provides an overview of computer image processing, including the estimation of performance and the development of algorithms. It explores the challenges faced in GPU accelerated remote sensing processing and the impact of changing conditions on program efficiency. The presentation also discusses the literature review, research questions, significance of the problem, aims of the research, research strategy and methodology, data gathering and analysis techniques, and the rigor, validity, reliability, and ethics of the research. It concludes with a research schedule and a list of references.

COMPUTER
IMAGE
PROCESSING

BACKGROUND TO THE PROBLEM
■ estimating the performance of image processing programs
■ The digital image processing has been considered as the vast research area
that includes both the pure theoretical practices and the works dedicated for
solving any specific problem
■ The operation of the GPU accelerated remote sensing processing is
significantly troublesome
■ the parallel strategies as well as the codes and algorithms for the GPU
accelerated remote sensing image processing could be commonly not
reusable
■ when changing the conditions or specific features of the objects, the
efficiency of the programs is sharply decreased.

LITERATURE REVIEW
■ GPU computing utilises the GPU as the co-
processor for accelerating the CPUs among all
the common- intention technical as well as the
engineering computing (Chen, 2015)
■ CPU comprise of numerous cores, while GPU
comprises of hundreds of ones (Damiand &
Lienhardt, 2014).
■ The GPU general computing is controlled by
programmability of the respective hardware
(Mori et al., 2018).
■ The CUDA parallel computing purpose
executing on the GPU could be referred as the
kernel or the (kernel function) (Russ, 2016)
■ Some research has been carried for increasing
the GPU algorithm library within these high-
level language (Sonka, Hlavac & Boyle, 2014).
■ Groups of the symmetric multi-processors
(SMP) are prevalent in the high performance
computing (Tanimoto, 2014)
■ .

RESEARCH QUESTION
■ What is the best programming template for the massive remote
sensing image processing?
■ How can the quality of the images from the various remote sensing
devices be enhanced and improved for better understanding?
■ How does enhanced image processing helps in the modern era and
how it can be effectively utilised by the humans?
■ What are the most appropriate algorithms that could be developed for
the enhancing of any kind of image?

SIGNIFICANCE OF THE
PROBLEM
■ It is clear that with the development of the
modern era image processing needs, there
is the major requirement of extracting the
best quality image for using in various
aspects.
■ The problem of not having the suitable
methods for the extraction of the images
leads to the lack of gaining the effective
visual data for analysis in various sectors.
■ The lack of appropriate algorithm for the
gaining of valuable images that could be
utilised for the analysis plays a significant
role in this research.
■ The image processing systems without the
improved and the modern and updated
technology could not capture the suitable
image data for analysis.

AIMS OF THE RESEARCH
■ The aim of this research is to determine which algorithms is suitable
for processing the computer images
■ This research aims to provide the generic idea about the working of
the computer image processing system and then propose the most
appropriate solution for the problems faced while image processing.

RESEARCH STRATEGY,
APPROACH AND
METHODOLOGY■ The research methodology that has been applied are the analysis of
the various kinds of algorithms for the determining of the accurate
algorithm and the system for gaining the enhanced images for the
analysis.
■ For the boundary detection problem, a number of wellknown program
mplementations of the algorithms based on the methods of the earlier
researchers
■ The method that is based on the simulation of the neurophysiological
processes of visual information processing is utilised for gaining the
solution for some of the problems associated with the image
processing

DATA GATHERING AND
ANALYSIS TECHNIQUES
■ The legitimacy of template in the GPU parallel program could be confirmed
by executing the GPU parallel trials of the three characteristic algorithms on
basis of the template.
■ Several algorithms of remote sensing image processing are applied on the
uppermost section of templates
■ In grouping of the logarithmic transformation, image inversion, Point
operations and the NDVI calculation have been nominated
■ In grouping of local operations, Gaussian filtering image and rotation have
been chosen
■ In grouping of global operations, and Fourier transform, and histogram
algorithm have been chosen

RIGOR, VALIDITY, RELIABILITY
AND ETHIC
■ The ethics could be applied in the research as it deals with the
analysis of the computer images that might sometime be of
significant importance to any other individual
■ The gaining of the images was done by the permission from the
proper government authorisation and the individuals who were
selected for the analysis of the proper technique for computer image
processing was done with the consent of every individual
■ The research is valid all through the complete cycle and there are no
violation of the ethics of the individuals

RESEARCH SCHEDULE
WBS Task Name Duration Start Finish Predecessors
0 Gnatt chart 119 days Mon 22-04-19 Thu 03-10-19
1 Research initiation phase 28 days Mon 22-04-19 Wed 29-05-19
1.1 Process identification 4 days Mon 22-04-19 Thu 25-04-19
1.2 Suitable employee identification 14 days Fri 26-04-19 Wed 15-05-19 2
1.3 Evaluation of standards and policies 5 days Fri 26-04-19 Thu 02-05-19 2
1.4 Justification of need of research 7 days Thu 16-05-19 Fri 24-05-19 4,3
1.5 Execute the feasibility study 3 days Mon 27-05-19 Wed 29-05-19 5
2 Planning phase 43 days Mon 27-05-19 Wed 24-07-19
2.1 Define the scope statement 4 days Mon 27-05-19 Thu 30-05-19 5
2.2 Divide the large deliverables 10 days Fri 31-05-19 Thu 13-06-19 8
2.3 Identification of required tasks for the production of deliverables 16 days Fri 14-06-19 Fri 05-07-19 9
2.4 Defining of the duration of the tasks 3 days Wed 17-07-19 Fri 19-07-19 10
2.5 Forecast the cost 3 days Mon 22-07-19 Wed 24-07-19 11
3 execution phase 14 days Thu 25-07-19 Tue 13-08-19
3.1 follow the created plan 2 days Thu 25-07-19 Fri 26-07-19 12
3.2 Assigning of the tasks to the team members 2 days Mon 29-07-19 Tue 30-07-19 14
3.3 Monitor the process 3 days Wed 31-07-19 Fri 02-08-19 15
3.4 Management of the contract that is secured in the project 7 days Mon 05-08-19 Tue 13-08-19 16
4 monitor and control 15 days Wed 14-08-19 Tue 03-09-19
4.1 Measure the research progress 8 days Wed 14-08-19 Fri 23-08-19 17
4.2 Measure the quality of the deliverables 7 days Mon 26-08-19 Tue 03-09-19 19
5 project closure 22 days Wed 04-09-19 Thu 03-10-19
5.1 Ensure completion of the research deliverables 8 days Wed 04-09-19 Fri 13-09-19 20
5.2 Close out the outstanding contracts 7 days Mon 16-09-19 Tue 24-09-19 22
5.3 Archive all paperwork 7 days Wed 25-09-19 Thu 03-10-19 23

REFERENCES
■ Chen, C. H. (2015). Handbook of pattern recognition and computer vision. World
Scientific.
■ Damiand, G., & Lienhardt, P. (2014). Combinatorial maps: efficient data structures for
computer graphics and image processing. CRC Press.
■ Mori, K. I., Kidode, M., Shinoda, H., & Asada, H. (2018). Design of a Local Parallel
Pattern Processor for Image Processing. In Special Computer Architectures for Pattern
Processing (pp. 197-210). CRC Press.
■ Russ, J. C. (2016). The image processing handbook. CRC press.
■ Sonka, M., Hlavac, V., & Boyle, R. (2014). Image processing, analysis, and machine
vision. Cengage Learning.
■ Tanimoto, S. (Ed.). (2014). Structured computer vision: machine perception through
hierarchical computation structures. Elsevier.