Week 3

Outline of the first version of fish eye pipeline.

  1.    convert color image to grey scale
  2.    histogram correction to help even out colour
  3.    threshold image to binary
  4.    extract contours from thresholded image
  5.    filter contours to get round contours
  6.    filter contours to get largest circle

Base on the above design, I create a pipeline and tried different set of parameter combination to make it work. For example, during the third step of the pipeline, threshold image to binary, I have tried threshold value from 50 to 150, and select the best to get the fish eye contour.

After that, filter out the noise to reduce number of contour identified by setting a threshold of contour circularity and contour size.

And finally, I finished the first version of fish eye pipeline and tested on the trevally fish image. And the result is good, I think!

Week 2

For the second week of my project, my goal is get the program running on my computer and have a basic design of my pipeline.

The program I am using is designed to read in the fish images as the input date and calculate or extract several features of the fish in the image such as: calculate fish contour, locate fish lip location. All calculated data will be written to different json file for further use. And base on the purpose or species of the fish, different pipeline will be used.

So, for testing the program on my computer, I have used a single image from trevally data set. And selected the trevally pipeline which we calculate the contour and extract the feature point of the fish. It also able to draw the calculated data on the image as requested. After playing with the program several days, I am able to get it running.

During this week, I have also learn the basic Python Syntax and the basic function of OpenCV. At the same time, I am trying to under how the program flow and how different function works under different pipeline.

Also, for my part of the project is develop a pipeline which can included in the program to extract the fish eye contour from the image. For the first stage of the design, it included the following steps:

  1. Gray scale the input image
  2. Histogram correction
  3. Threshold image to binary
  4. Extract contours
  5. Filter round contours
  6. Filter to get largest circle

Week 1

For the weeks of the project, I went to Plant and food Research (PFR) and meet David who is my supervisor in PRF. He shown me what I should do in the project and help me get the seat and computer in PFR.

The project I am doing is using a program which is written by Python. So, I need to install all the required software and library to get it running on my computer. Also GitHub is also used for version control. 

So, I have install all necessary software as below: 

Install Github and get the repository

Install Anaconda

Install all package: sys, pytest, cv2, scipy, skimage, sklearn, sqlitedict, pandas, numpy, xlrd

After all the installation, I try to get understand how the program works and do some path setting in PyCharm. This included the environment setting and a json document setting which contain the input and output data direction path. With the help of Jacky who is a programmer based in Auckland office, I got everything ready for the text run next week.

PRJ 702 2020 04 10

To overcome the problem mentioned last week, I Used Canny to outline the edge detection function which was implemented in OpenCV and successfully identified the fish eye contour for all 15 Trevally data.Typical steps of using canny required a 5×5 gaussian filter to smooth the image first. Then, using the canny edge detection function to outline all the edges in the image.

The image showed the result of canny function, able to detect the edge of the fish eye. 

After that, same as the previous step, perform a binary threshold of the image and substrate the canny edge result to add the edge. The image below shown the edge added to the image with binary threshold processed.

The image showed the image with binary threshold combined with canny edge detection result

After another filter and contour extraction, the pipeline is able to get a single fish eye contour on trevally data

Final result of the fish eye extraction pipeline.

Next step is testing the performance of the fish-eye pipeline on snapper data.

https://docs.opencv.org/trunk/da/d22/tutorial_py_canny.html

Reflection on the course

Thank you so much for Lars teaching us some many thing during this term. I have learned a lot of things such as how design a research project, how to determine the scope of the research and how to evaluate the information from different sources such as internet or library to help us understand the problem we are facing.

I am really enjoying the first few lectures, learning about what is research, what is true or truth. If I did not take this course, I will never think about these questions in my life. And I believe having this knowledge is the fundamental knowledge to understand the research approach.

After the basic background of the research, this course helps me a lot to find my research interest. Lars teach us how to find the academic paper, using referencing tools and guide us how to narrow down our research topic step by step.

And finally, we can understand our research interest in information technology area and able to apply what we have learned to launch a research project.

Thanks a lot!!!

Area of research interest peer review

I read Marcellus’s research interest and find out that he is going to do an interesting research project, virtual reality therapy. I think this is a very new way to use technology for treatment in various psychological disorders. But after reading Marcellus’s blog, it turns out the therapy has been conducted since 1990s. And as the technology improved a lot during these yeas and the equipment like VR headset being popular, the market would have a huge demand of these kind of VR app specific in research area.

 I believe Marcellus spend a lot of time on this topic and demonstrated knowledge skills and abilities to conduct any of the project. He listed out several possible project in these areas and making a choice of using personal smartphone is a very smart decision. Because smartphone is a popular daily life equipment, we can obtain it very easily. This can make sure the access of a suitable equipment during the development state. Also, he has high level of programming skill in C# and able to unity3D to develop games which already make a start of the project.

I am really looking forward to seeing his project using in the research area or even in the hospital.

ETHICAL BEHAVIOUR

Examples: You find some good images on the web that would look good on your site.  You copy and paste them to your blog.

Almost everybody needs to learn from the other, also all of us looking for good material for our blog. If we found something good, share it with other people is a good idea. If we have a good reference and give the original author a good credit, putting good images in our website or blog is acceptable.

Examples: You are in a hurry to finish an assignment and you find the perfect explanation of a difficult concept on Wikipedia. You decide to copy it into your work.

I think copying thing from other is not a ethical behavior in this example. Because the motive of coping this is only because you don’t have time to finish not to share the good idea to make the world better.

Completing the assignment is our responsibility, we have to spend to do the research, structure the knowledge for what we learn and do the assignment. Not copy it from the other person or Wikipedia.

Examples: You are a medical researcher and you think you have discovered a new drug that cures lung cancer.  It worked well on mice although a few of them died of heart problems. You want to test it on people.  You recruit 100 lung cancer patients into your treatment program. You don’t tell them about the new drug.  You treat 50 patients with usual drugs and 50 with the new drug to see which is better.

I think we don’t need to tell the patient which drug they are having, and this is an ethical behavior. Because if we let the patient know which kind of drug they are having. Some people would think new drug is better than the old one and make the research not accurate.  

Examples: You are doing your PhD about drug abuse among students.  You interview 30 students about their drug use. You discover that one of the students is actually a fairly big time dealer and you report him to the police.

As we discover a drug dealer, we must report him to the police no matter we are in a research project or not. So to perform this kind of research, we have to prevent recording the identity of any subject. Otherwise we don’t have any way to do this kind of research.

Examples: You are a researcher looking at the effect of violent computer games on children.  You recruit 20 children into your study. Over a month you regularly show them images of violence to see if it has an adverse effect on their behavior.  One of the children becomes quite distressed each time and so you stop showing her images and drop her from the study.

Showing violent computer game to children is an unethical behavior. Because it could cause an unpredictable result to the children. To perform this kind research, we better collect sample from the child already playing violent games, but not give them the violent game.

Examples: You want to research how easy it is to hack into your organization’s computer system by persuading people to divulge login and password details.  You recruit a small team to ring up key people in the organization and persuade them to give either their own or their boss’s details.

To do this kind of IT security research or testing, we must define the scope of the project in advance. We need to have a list of procedure to control the researcher activity such as what level of hacking is acceptable. And show this to all people amount the organization to let them know what will happen and collect they approval. As the research process, all the research mush stick to the plan, the leader of the project must make user everyone has fellow the procedure.

Examples: You are doing an initial research in area of a town where may bars and pubs are located, to estimate the level of problematic social behavior in the area. your research is independent from the police, because you want to observe their behavior as well. You observe both abusive and violent behavior.

I think we need to prevent any abusive and violent behavior. So, if we don’t report that is an unethical behavior. Before starting the observation, we need to define a bottom line in advance. Such as if we see getting bitten, we must report to the police. This can allow us to finish our research and no body will get serious injure.

RES701 Broad Research/Project Interest Area

  • Short description of the area of IT that interests you (explain it to me!)
    • Data science
    • Machine Learning
    • biomedical mining
  • Why is it interesting to you
    • As the thought put of NGS output increase a lot and tone of the data is available and I have some kind biological background, so I hope I can do something like Data Analyst vs. Data Scientist or
    • Do something with machine learning to use the data
  • Three things you know about it
    • The data size in bioinformatics is increasing dramatically in the recent years.
    • Availability of high volume of data is helpful for more accurate analytics, particularly in a highly sensitive field of research like bioinformatics
    • In order to tackle these challenges of big data in bioinformatics, cloud computing technologies have been used, with a lot of success.
  • Three things you believe about it
    • NGS data is the future of biological research tool.
    • Application of Machine learning in bioinformatics helps to deal with the big data.
    • Problems in bioinformatics facing huge and fast growing dataset.
  • Three things you don’t know about it  
    • I don’t know the traditional way of analysis dig data
    • The limitation of the traditional way of analysis biological data.
    • The application of mechine learning in bioinformatic
  • Would you rather: Do something, Research it or Do both
    • I want to establish or combine different dataset for clinical use.
  • Go and see if you can find a few resources about it – (post URL) forum? facebook group? linkedin group? or a good blog? Or even academic articles 🙂 
    • Kashyap(2015)
  • AND – APA reference them in your writeup. 

Kashyap, H., Ahmed, H. A., Hoque, N., Roy, S., & Bhattacharyya, D. K. (2015). Big Data Analytics in Bioinformatics: A Machine Learning Perspective. ArXiv:1506.05101 [Cs]. Retrieved from http://arxiv.org/abs/1506.05101

AREAS OF INTEREST

The IT area/subject I have most enjoyed is programming and data base

The IT area/subject I least enjoyed is security

The IT area/subject I was most interested in is data analysis

The one IT thing I never want to have to do again is security

I chose to study IT because I  want to have a job related to both biology and IT

If I couldn’t study IT I would study Biology

When I was a kid I wanted to be Biologist

One IT thing I would like to know more about is database.

Design a site like this with WordPress.com
Get started