The total marks for the individual and group components of the project , as well as comments, have been released. You can access your marks by logging in to your account and typing:
9517 classrun -sturec
The WebCMS3 system updates overnight so you may not be able to access the marks through WebCMS3 until tomorrow. To see your marks on WebCMS3:
PLEASE NOTE: no emails regarding marks will be answered until after the holiday period.
The exam for COMP9517 is this Friday at 8:45am. If you have any questions about material covered in the course the lecturers and tutors are available for consultation this week at the times specified on the course website under Consultations. More times have been opened up, so please check the timetable. They are there to help you, so please make use of this time.
Dear students, this is a reminder to fill in the myExperience survey for COMP9517 if you have not already done so- you can use the week 10 lecture hour to fill it in.
You can login on your laptop, tablet or phone and:
There is no lecture and no lab on Friday 22nd. You may use the time to finalise your reports.
Please contcat your tutor directly if you have any questions on the report. If you would like to
meet with your tutor, you should arrange it by contacting them.
Lecturers/tutors will be available for consultation about the exam at the times listed under Consultations on the course website. Please check this page periodically as it will be updated.
Tutors will be available for consultation on the project TODAY from 11am-1pm and 3-5pm in K17 Room 403 on the 4th floor. Please make use of this time to make sure you are on track for the demonstrations on Tuesday and final submission on Friday.
The submission links for the individual and group components of the project are now open.
Please compress all your files (report and code) into a .zip file for upload. That is the only file format that will be accepted. The file size limit is 20Mb.
There will be additional consultation times available on Monday 18th November where you may get help with the project, ahead of the demonstrations on Tuesday.
Tutors will be available from
11am - 1pm
and from
3pm - 5pm
in
K17 Room 403
on the 4th floor.
Please take advantage of these sessions so that you are well prepared for Tuesday and for the project submissions on Friday.
Some sample exam questions and answers are now available on the Course Website under Course Work->Assessments->Exam.
Instructions for the final project demo are now available on the course website under Course Work->Assessements->Project->Final Project Demo Instructions - 19T3.
In Friday's lecture 3-4 PM, I shall give a quick recap of project demo and final exam.
Dr Gihan Samarasinghe will give a talk on Computer Vision for Diagnosis of Prostate Cancer and Radiation Therapy Planning .
The schedule for the project demonstrations in Week 10 has been updated due to the unavailability of a tutor. The new schedule is available on the course website under Course Work->Assessments->Project->Project Demo Schedule - 19T3. All projects will be marked by one tutor and one academic who are familiar with the project.
The schedule for the project demonstrations in Week 10 is now available on the course website under Course Work->Assessments->Project->Project Demo Schedule - 19T3. ALL members of each group MUST attend the demonstration. Please aim to arrive 10 mins prior to your scheduled time to ensure that the program flows smoothly. Demonstrations will be strictly limited to 15 mins.
In week 9, we have a series of presentations on Computer Vision applications, details of Tuesday's talk appear below.
Information on Friday's talk will be made available soon.
No slides will be uploaded, as this is based on research work and/or industry product development. This material is also not directly examinable, and we hope that they will act as inspiration and motivation.
Tuesday Nov 12th, 3-4 PM
Kane Walter
, Deep Learning Query-be-example system for remote sensing imagery
Danni Ovens , Automated Grading of Refractive Error in Fundus Images using Deep Learning
Tuesday Nov 12th, 4-5 PM
Invited Talk by John Pearce, RocketBoots Sydney
Title : Here Come the Autonomous Paparazzi
It takes one to know one, especially when it comes to face recognition. While face recognition algorithms get plenty of attention, getting a useable enrollment image of a person for face recognition is a key prerequisite to being able to use face recognition in real world, high traffic situations.
In this talk, John Pearce will describe the process RocketBoot’s face recognition solution uses to autonomously collect usable enrollment images for face recognition using a combination of face detection networks, clustering and more traditional computer vision algorithms.
Speaker Bio : John Pearce completed a Bachelor of Software Engineering (Honours first class) at UNSW in 2011. As a Software Engineer at RocketBoots , John loves to code and see stuff "just work", attempts to resist his perfectionist urges to refactor other people's code, in between office disputes over which secret Avengers persona each team member hides and whether Coke or Pepsi taste better (or perhaps even the same...). His personal life is marked by being jumped upon by many small humans and hanging out with his wife.
The T3 2019 myExperience survey is now open.
You can complete your survey via:
Please take the time to complete the survey. Your constructive feedback is highly valued and appreciated.
The marks for Labs 4 & 5 and the total lab mark have been released. Late penalties, if any, have been applied to the total mark.
You can access your marks by logging in to your account and typing:
9517 classrun -sturec
The WebCMS3 system updates overnight so you may not be able to access the marks through WebCMS3 until tomorrow. To see your marks on WebCMS3:
The slides for this week's lectures are now available on the course website under Course Work->Lectures->Week 8.
The submission link for Lab 5 is now open. It will close at 23:59 tonight.
No late submissions will be accepted!
Please compress all your files into a .zip file for upload. That is the only file format that will be accepted. The file size limit is 5Mb.
The Pytorch tutorial is also available on the course website under Course Work->Lab Activities->Week 7.
Lab 5 is now available on the Course website under Course Work->Lab Activities->Week 7 and is due on Fri 1st November at 23:59 (midnight). The submission link will open at 4pm on Friday.
The slides for this week's lectures are now available on the course website under Course Work->Lectures->Week 7->Week 7 - Deep Learning for Computer Vision (19T3).
If anyone mistakenly picked up a black, leather laptop cover in the lecture room today, it belongs to one of the tutors. Could you please email cs9517@cse.unsw.edu.au to arrange its return. Thanks.
The marking criteria for the project have been released and can be found on the course website under Course Work->Assessments->Project.
As already discussed in class last Friday:
This week ( week 6) is flexible week
• No lectures on Tuesday and Friday
• No formal laboratory session on Friday
Use the week to start work on your project
• contact your assigned tutor via email and arrange a group meeting with them
• they will help to get started and answer any questions
• Tutors will keep notes on meetings, so should you
The submission link for Lab 4 is now open. It will close at 23:59 tonight.
No late submissions will be accepted!
The slides for today's lecture have been updated and are available on the course website under Course Work->Lectures->Week 5.
The marks for Assignment 1 and Lab 3 have been released. Late penalties, if any, have been applied to the assignment mark, and will be applied to the lab marks at the time of grading.
You can access your marks by logging in to your account and typing:
9517 classrun -sturec
The WebCMS3 system updates overnight so you may not be able to access the marks through WebCMS3 until tomorrow. To see your marks on WebCMS3:
The slides for Friday's lecture are now available on the course website under Course Work->Lectures->Week 5->Week 5 - Pattern Recognition - Part 2 (19T3).
The presentation on the project given in yesterday's lecture is now available on the Course Website under Course Work->Assessments->Project->Project Presentation (19T3).
A list of the tutors that have been assigned to mentor each project group has been released and can be found on the course website under Course Work->Assessments->Project->Project Mentors.
From Week 6 on you must meet with your tutor at least once a week at a mutually convenient time. Tutors will be keeping a record of your meetings.
The group named "I have no idea what to call it so just let it be then" has only 3 students. We need someone to swap out of one of the groups of 5 students into this group. Please contact the students in that group if you wish to do that.
The marks for Lab 1 and 2 have been released. Please note that late penaltie, if any, will be applied at the time of grading. You can access your marks by logging in to your account and typing:
9517 classrun -sturec
The WebCMS3 system updates overnight so you may not be able to access the marks through WebCMS3 until tomorrow. To see your marks on WebCMS3:
The slides for today's lecture have been updated and can be found on the course website under Course Work->Lectures->Week 5.
If you have not yet joined a project group, there are several groups that have less than 5 students. Please try and join one of those groups rather than starting another new group.
The specification and data files for the project have been released and can be found on the course website under Course Work->Assessments->Project.
The project specs will be discussed in class tomorrow.
The slides for Tuesday's lecture are now available on the course website under Course Work->Lectures->Week 5.
This week's lab is also available under Course Work->Lab Activities->Week 5 and is due on Fri 18th October at 23:59 (midnight). The submission link will open at 4pm on Friday.
Just a reminder that we will be accepting a range of solutions to the water shedding algorithm on the strawberry.
The main criteria we are looking for for this lab is that the meanshift algorithm works successfully for the strawberry and the watershed algorithm works successfully for shapes.
For those that are curious as to how to get the watershed algorithm to identify the whole strawberry (which was shown in the lab) there are two main methods to achieving this.
The first, is to use a binary threshold on the image. The second is to remove the "labels" parameter from the peak_local_max function. This parameter essentially tells the algorithm to include the pixel intensity as well as the distance to the background. With an image like the strawberry, this causes a rainbow looking effect as there are so many slight colour changes caused by the shadows, the spots and the stem. Whereas just using the distance to the background to segment the image correctly identifies the whole strawberry as one object.
Please remember to form groups of 5 for the project and register your groups on the course website . The deadline will be extended to Monday 14th October. After that time, if you have not formed a group you will be allocated to one. The permission of the LIC is required for smaller groups.
To register a group, go to Groups and click on Create (top right). Give your group a name and make sure you select " Group Project " from the Group Type drop-down . You can then add other students to your group by clicking on the add button. Alternatively you can click on the Join button beside the name of an existing group to join it. Please create your group name from your own names or initials. Inappropriate group names will not be accepted.
The submission link for Lab 3 is now open. It will close at 23:59 tonight.
No late submissions will be accepted!
The slides for today's lecture are now available on the course website under Course Work->Lectures->Week 4
Lab 3 is now available on the Course website under Course Work->Lab Activities->Week 4 and is due on Fri 11th October at 23:59 (midnight). The submission link will open at 4pm on Friday.
The solutions to Lab 2 are now available on the course website under Course Work->Lab Activities->Week 3.
The lecture slides for today's lecture are now available on the course website under Course Work->Lectures->Week 4->Week 4 - Segmentation - Part 1 (19T3).
The project spec will be released this week. The project will be carried out in groups of 5 people . Please register your project groups on the course website by Friday 11th October (week 4).
To register a group, go to Groups and click on Create (top right). Give your group a name and make sure you select " Group Project " from the Group Type drop-down . You can then add other students to your group by clicking on the add button. Alternatively you can click on the Join button beside the name of an existing group to join it. Please create your group name from your own names or initials. Inappropriate group names will not be accepted.
If you cannot form a group of 5 you will be allocated to a group. The permission of the LIC is required for smaller groups.
The submission link for Lab 2 is now open. It will close at 23:59 tonight.
No late submissions will be accepted!
The lecture slides for Week 3 are now available on the course website under Course Work->Lectures->Week 3->Week 3 - Feature Representation (19T3).
Lab 2 is now available on the Course website Course Work->Lab Activities->Week 3 and is due on Fri 4th October at 23:59 (midnight). The submission link will open at 4pm on Friday.
The solutions to Lab 1 are also available on the course website under Course Work->Lab Activities->Week 2.
An error in the Laplacian mask in Question 4 has been corrected. The correct central value is 4 (not 8 as in previous version).
Please use this in your submitted work.
Also confirms that just one Laplacian image should be submitted as the solution to Question 4.
The submission link for Lab 1 is now open. It will close at 23:59 tonight.
No late submissions will be accepted!
Please compress all your files into a .zip file for upload. That is the only file format that will be accepted. The file size limit is 5Mb.
The marking criteria for the assignment have been released, look under Assessment ---> Assignment 1
Submission for Lab 1 will open at 4pm on Friday- this will be the practice for all future labs too.
The link is not accessible at the moment.
You can now submit your files for Assignment 1 via the course website: Course Work->Assessments->Assignment 1->Assignment 1 Submission (19T3). Instructions for the submission are also given at this link. Deadline: Mon Oct 7th 23:59:59.
The first lab for this course will be held on Friday Sept 27th, from 4-5 PM, during regular class hour , in the scheduled class room .
Tutors will be available to assist you while solving lab problems.
There is a single assessable question. You may submit your answer to it online on Friday Sept 27th, by 23:59:59 hours.
These have been uploaded.
The specifications for the assignment and the sample images have been released under
Assessments --> Assignment 1
The second lecture for COMP 9517 will be held this Friday from 3-5 PM.
The venue is ElecEngG23, which is different from the Tuesday venue.
The lectures slides for the SECOND lecture of Week 1 are now available on the course website under Course Work->Lectures->Week 1-> Week 1 - Image Processing Part 1 (19T3).
The general practice would be to make draft slides available a day before the lectures, and upload an updated version soon AFTER the lecture, in case there are any updates.
The lectures slides for the first lecture of Week 1 are now available on the course website under Course Work->Lectures->Week 1->Week 1 - Introduction (19T3).
The first lecture for COMP 9517 this term will take place on Tuesday, Sept 17th, 3-5 PM in Colombo Theatre A.
See you all there.
Dear COMP 9517 students
Welcome to COMP 9517. The draft course outline is available via the link on the left panel.
Classes will start on Tuesday week 1. See you all then.
Best wishes
Sowmya and the COMP9517 team