Team Project

The goal of team projects is to give students an opportunity to experience team work, to practice technical report writing, to practice presentation skills, and to gain new knowledge in the field of neural networks.

Contents

  1. Team Formation
  2. Project summary
  3. Final Report
  4. Deadlines
  5. List of Project Teams
  6. List of Suggested Projects

Team Formation

At the start of semester all students enrolled in the course "Neural Networks" should form project teams consisting of at least four and of no more than six students per team.

You may use the forum available at http://www.fer.unizg.hr/en/course/neunet_a/project/forum to form teams and to select project topics.After a project team is formed team members should appoint a team leader. Selected team leader should then inform teaching assistants and lecturers via e-mail about the formed project team and about selected project title.

Project Summary

Each project team should prepare a short project summary which contains the following information:

  1. project title,
  2. project topic and short description (an abstract between 100 and 200 words),
  3. project breakdown into tasks (What needs to be done?),
  4. project results and deliverables (What is the end product?),
  5. team member assignments and tentative time schedule.

Final Report

At the end of semester each project team has to submit a final report.

The final report is comprised of:

  1. a written technical report;
  2. a presentation to other enrolled studens and to the lecturers;
  3. of all data and developed program code which are submitted directly to the lecturers.

Written Technical Report

A written technical report should be submitted as a PDF document via e-mail.

Please use the A4 IEEE Conference Template as a template for your written technical report.

IEEE Conference Templates are available at https://www.ieee.org/conferences_events/conferences/publishing/templates.html.

We recommend the following structure of the written technical report:

  1. a short introduction including problem statement and motivation;
  2. an overview of the existing work, i.e. a short literature review;
  3. a description of the solution implemented by the project team;
  4. a description of experimental results;
  5. a discussion comparing your results with previous results from the literature (if any);
  6. a conclusion; and
  7. a list of references.

We also recommend you do not overload your technical report with developed program code as it is submitted separately.

Project Presentation

At the end of the semester each project team will present their results. Presentation aids are computer and projector. This presentations will be held during regular lecture hours in the week which preceeds the final exams. Exact time alloted to each team will be announced during the week preceeding the presentations.

Presentation should be submitted as a PDF document via e-mail.

Attendance is mandatory: all enrolled students and all team members must be present during final project presentations.

Source Code and Data

Developed program code and used data should be submitted directly to the lecturers by the project leader.

Deadlines

The following deadlines are firm:

  • Formation of project teams: 09. 10. 2020.
  • Submission of project summaries: 16. 10. 2020.
  • Consultations with professor on project's progress: until 13. 11. 2020.
  • Final report: 11. 1. 2021.

Missing any of the aforementioned deadlines may yield negative penalty points unless there are extenuating circumstances.

List of Project Teams

No. Project Title Team Members Presentation time Adversary team no.
1 Traffic sign detection using CNNs Marko Puk, Lovro Levak, Goran Petrak, Filip Radović, Antonio Pavliš 13.1.2021. at 16.00h 3
2 Emotion recognition Franko Šikić, Emilia Szymanska, Igor Zielinski, Marko Šandrk, Jinyuan Li 13.1.2021. at 16.15h 8
3 GANs for generating different face expressions Helena Čeović, Andrea Omićević, Josipa Lipovac, Antun Magdić, Ante Žužul 13.1.2021. at 16.30h 2
4 Real-time face blurring for minors Filip Karlo Došilović, Antonio Gauta, Lara Lokin, Domagoj Pavlović, Ana Puljčan 13.1.2021. at 16.45h 10
5 Pedestrian detection using CNNs David Dukić, Domagoj Jakšić, Ante Čurić, Šimun Krmek 13.1.2021. at 17.00h 7
6 Football player detection on the football field Ivona Farkaš, Veronika Knezović, Ante Gazibarić, Jelena Nemčić 13.1.2021. at 17.15h 5
7 Neural networks for unstructured datasets processing (DeepSets and/or PointNet) Zvonimir Sučić, Marko Pisačić, Marin Delić, Bruno Sačarić, Filip Milić 13.1.2021. at 17.30h 6
8 Recognizing the gender and the age of a person Gabriela Mišković, Ana Leventić, Marica Mikulić, Josip Pečarić, Luka Kušina 13.1.2021. at 17.45h 11
9 Layer and fluid segmentation in OCT retina images Giovanni Camisa, Filip Oreč, David Gorup, Tomislav Grdenić, Sandi Letić 20.1.2021. at 16.00h 4
10 Face mask detection using CNNs Tin Oroz, Petar Mihaljević, Matija Roglić, Martin Pavić, Marin Mrčela 20.1.2021. at 16.15h 9
11 Different cloud formations classification using satellite images Anton Moysey, Maksimilijan Novak, Ivana Cvitanović, Arnaud Nathan Gueze 20.1.2021. at 16.30h 1

List of Suggested Projects

Here is a list of suggested project topics and lecturer whom you may contact for more information if you are interested in a particular topic:

  • Layer and fluid segmentation in OCT retina images
  • Traffic sign detection using CNNs
  • GANs for generating different face expressions
  • Emotion recognition
  • Pedestrian detection using CNNs
  • Football player detection on the football field
  • Different cloud formations classification using satellite images
  • Diabetic retinopathy stage detection using fundus images
  • Neural networks for unstructured datasets processing (DeepSets and/or PointNet)
  • Approximate Bundle Block Adjustment problem solving (Dense Bundle Adjustment Network)
  • Virtual character animation using the OpenPose neural network

Repository