IBM-Cortex Hackathon

MIIA Deep Learning Hackathon – 16 September 2017 from 9am-8:30pm @ The DIZ, 111 Smit Street, Braamfontein, Johannesburg

MIIA AI/Data Science Hackathon – 16 September 2017

We have also scheduled our first Data Science Hackathon sponsored by IBM and Cortex Logic at The Diz, 111 Smit Street, Johannesburg on Saturday 16 September 2017 from 9am-8:30pm. As this hackathon follows the Deep Learning Indaba week, the theme for the hackathon is also Deep Learning and the event is called Deep Learning Hackathon.

[We are also working behind the scenes on a Data Science Hackathon in Cape Town. Anyone interested to participate or sponsor these events are welcome to email us at Please also specify the location that is relevant for you.]

SponsorsIBM, Machine Intelligence Institute of AfricaCortex Logic, The DIZ

Venue: The DIZ, 3rd Floor, 111 Smit Street, Braamfontein, Johannesburg (corner of Smit and Eendracht street). Parking on Juta and Eendracht street.


Challenge: Computer Vision problem – having a closer look at data related to a major problem affecting transport in South Africa and costing the country in the order hundreds of millions of Rands…

Announcement of Challenge: The computer vision challenge will be revealed at MIIA Event on the 15th of September at the same venue as the hackathon (Quicket:; Meetup: We’ll also post this on the MIIA website (Events page:

9am-5pm: Working on Challenge

  • 9-9:10am: Kick-off and formation of teams (can also be individual participants)
  • 1pm: Lunch
  • 5pm: Participants must submit predictions on test set

5-6pm: Working on Presentations

  • 18:30: Drinks, snacks + working on presentations

7-7:15pm: MIIA, Introductions to Panel of Judges and Sponsors

  • 7-7:05pm: MIIA
  • 7:05-7:10pm: Introduction to panel of judges
  • 7:10-7:15pm:  IBM as main sponsor

7:15-8:15pm: Team/participant presentations

  • We aim for maximum of 2 minutes per team/participant presentation

8:15-8:30pm: Presentation of Prizes:

  • Judges present prizes:
    • for most accurate models
    • for best presentation (fun prize)

To register on Quicket: 


MIIA Events page:


MIIA Event on Friday, 15 September 2017 @ The DIZ,  111 Smit Street, Braamfontein, Johannesburg at 5:30pm

Deep Learning and its Applications

SponsorsIBM, Machine Intelligence Institute of AfricaCortex LogiceXe-ML, The DIZ

Venue: The DIZ, 3rd Floor, 111 Smit Street, Braamfontein, Johannesburg (corner of Smit and Eendracht street). Parking on Juta and Eendracht street.


• “Introduction and MIIA update“, Dr Jacques Ludik, MIIA, Cortex Logic, Bennit.AI

• “Info session for Deep Learning Hackathon on 16 September“, Alex Conway, Founder & CTO, NumberBoost

• “Convolutional Neural Networks for Computer Vision Applications“, Alex Conway, Founder & CTO, NumberBoost

• “Conversational Systems: Beyond the Interface”, Chris Currin, PhD Computational Neuroscience candidate at UCT

• “Deep Learning and its Applications”, Dr Jacques Ludik, MIIA, Cortex Logic, Bennit.AI

• “Introduction to Bluemix“, IBM 

To register on Quicket: or


MIIA Events page:

Computer Vision Challenge

Training & Test Data and Python Notebooks

Stellenbosch University benchmark

Nienaber, S. ; Booysen, Marthinus J. ; Kroon, R. S.
34th Annual Southern African Transport Conference SATC 2015 – Theme – Transport: Working together to deliver – ‘SAKHA SONKE’, 6 to 9 July 2015, CSIR International Convention Centre, Pretoria, South Africa .

The original proceedings are available at 
Potholes are a nuisance, especially in the developing world, and can often result in vehicle damage or physical harm to the vehicle occupants. Drivers can be warned to take evasive action if potholes are detected in real-time. Moreover, their location can be logged and shared to aid other drivers and road maintenance agencies. This paper proposes a vehicle-based computer vision approach to identify potholes using a window-mounted camera. Existing literature on pothole detection uses either theoretically constructed pothole models or footage taken from advantageous vantage points at low speed, rather than footage taken from within a vehicle at speed. A distinguishing feature of the work presented in this paper is that a thorough exercise was performed to create an image library of actual and representative potholes under different conditions, and results are obtained using a part of this library. A model of potholes is constructed using the image library, which is used in an algorithmic approach that combines a road colour model with simple image processing techniques such as a Canny filter and contour detection. Using this approach, it was possible to detect potholes with a precision of 81.8% and recall of 74.4.%.


Model accuracy on Test Data (top 7 teams):

1. Jandre Marais – 85.58%  –> First Prize 

2. aiThenticate Team – 73.03% –> Second Prize

3. ThinkSmart Team – 73.00% –> Third Prize

4. Urban Team – 70.97%

5. ThinkData Team – 61.00%

6. HackOverflow Team – 58.06%

7. Shallow Learning Team – 57.51%


Best Presentation:

1. Shallow Learning Team

Winning Solution

Jandre Marais: 

Photos of the event

Model accuracy on Test Data:

1. Jandre Marais – 85.58%  –> First Prize 

2. aiThenticate Team – 73.03% –> Second Prize

3. ThinkSmart Team – 73.00% –> Third Prize

Best Presentation:

1. Shallow Learning Team

Loading posts...
Sort Gallery