De retour ici (back here)

I realize I would benefit from some space to blog some ideas, more or less random, and air some frustrations. So they may be some more content as I am also in the middle (already) of a new (third) 6-month sabbatical, trying to figure out what to do over the next medium-term (3-5 years).

Recent influential readings include Four thousand weeks by Oliver Burkeman and The pathless path by Paul Millerd. More to come.

TRB 2014 Activities

This TRB is very busy as I am co-organizing a workshop on Sunday afternoon, co-chairing the meeting of the subcommittee on surrogate safety measures and a co-author in 7 papers, 3 oral presentations, 4 posters (and 1 withdrawn by the student and first author):

Day Title Committee Sponsor
Sunday 1:30PM- 4:30PM
Workshop 162
Comparison of Surrogate Measures of Safety Extracted from Video Data
Nicolas Saunier, Polytechnique Montreal, Canada, presiding
Data and Information Technology, Hot Topic: Big Data, Safety and Human Factors
Monday 1:30PM- 3:15PM
Session 353
Advances in Geospatial Technology Applications in Transportation
James P. Hall, University of Illinois, Springfield, presiding
Lectern Session | Practice Ready Papers
Data and Information Technology
Monday 1:30PM- 3:15PM
Session 341
Roadway Lighting: Technologies, Visibility, and Safety
Wade Odell, Texas Department of Transportation, presiding
Lectern Session
Design, Safety and Human Factors
Monday 1:30PM- 3:15PM
Surrogate Measures of Safety Subcommittee, ANB20(3)
Andrew P. Tarko, Purdue University, presiding
Nicolas Saunier, Polytechnique Montreal, Canada, presiding
Data and Information Technology, Safety and Human Factors
Monday 2:00PM- 3:45PM
Session 380
Developments in Active Transportation: Meeting the Challenges of Demand for Data in Bicycle and Pedestrian Fields
Elizabeth Stolz, Sprinkle Consulting Inc, presiding
Poster Session
Data and Information Technology, Hot Topic: Big Data, Pedestrians and Bicyclists
Tuesday 8:30AM- 10:15AM
Session 527
Emerging Information Technology Advances in Transportation, 2013 [10]
Guohui Zhang, University of New Mexico, presiding
Poster Session
Data and Information Technology
Tuesday 10:45AM- 12:30PM
Session 562
Alternative Safety Performance Indicators: Advancing the Frontier
Andrew P. Tarko, Purdue University, presiding
Poster Session
Data and Information Technology, Safety and Human Factors
Wednesday 10:15AM- 12:00PM
Session 786
Roundabout Perception and Reality: Speed, Pedestrians, Capacity
Andrew Paul, Massachusetts Department of Transportation, presiding
Lectern Session | Practice Ready Papers
Design, Operations and Traffic Management, Safety and Human Factors

Public Video Dataset and TRB Workshop on the Comparison of Surrogate Measures of Safety Extracted from Video Data

Call for Participation – 2014 TRB Annual Meeting Workshop
Comparison of Surrogate Measures of Safety Extracted from Video Data

Sponsoring Committee:
Safety Data, Analysis and Evaluation (ANB20)
Co-Sponsoring Committee:
Artificial Intelligence and Advanced Computing Applications (ABJ70)

With the advent of powerful computer vision techniques, video data can be automatically analyzed for an increasing number of transportation applications, including for road safety diagnosis based on surrogate measures of safety that do not require to wait for accidents to happen. While several methods exist for different purposes and settings, few direct comparisons have been made and no guidelines exist to choose and adjust existing methods for a given application. An important reason is the lack of public datasets and of comparison of state of the art methods on tasks relevant for transportation applications (benchmarking). Public datasets and benchmarking are common in several scientific fields, most notably in computer vision (such as the series of IEEE International Workshop on Performance Evaluation of Tracking and Surveillance (PETS) , with 15 workshops from 2000 to 2013) but not in transportation. There is in fact a lack of public data relevant for transportation applications: some general computer vision datasets can be useful, but cover few real life transportation applications.

A group of researchers from the University of Lund, McGill University and Polytechnique Montréal have decided to create such a public video dataset, in particular for road user behaviour analysis and road safety diagnosis. We invite the researchers and practitioners to present the current state of their methods for video analysis, behaviour and safety diagnosis, whether previously published or not. We encourage the workshop participants to test their tools on the videos from the public dataset so that the performance of the different system can be directly compared.

The workshop will take place at the next TRB annual meeting on Sunday afternoon, January 12th 2014 (1:30pm- 4:30pm in Marriott, Madison A). Everyone is welcome to attend and participate to the discussions.

The information about the workshop, including the links to the datasets, is available on the following webpage:

For more information please contact Nicolas Saunier at or Aliaksei Laureshyn at

Organizers: Hakan Ardo and Aliaksei Laureshyn (Lund University), Luis Miranda-Moreno (McGill University), Nicolas Saunier (Polytechnique Montréal).

August: the video dataset is made public
extended deadline November 15th: submission of an abstract by email to the organizers describing the methods and results
November 25th: selection of participants and final workshop program

New Tools for Replicability

Finally some new posting on this blog. You may have read in The Economist and other places about the various problems with research and science. If you know me or have read my professional website or follow me on twitter, you know I am a big proponent of reproductibility, the capacity for anyone to replicate another researcher’s findings. I sometime feel lonely in the field of transportation. That is why I am happy when I find strong confirmation of the crucial aspect of reproductibility elsewhere. I will therefore quote from a previous post of that New Yorker blog a few sentences (emphasis added):

“Instead of, for example, rewarding scientists largely for the number of papers they publish—which credits quick, sloppy results that might not be reliable—we might reward scientists to a greater degree for producing solid, trustworthy research that other people are able to successfully replicate and then extend.”

Without replication, all results should be taken as preliminary.”

“The best science is cumulative, not just a list of fun results; as people push deeper, bad ideas that are invalid eventually crumble.”

That is why I push for open data, open source software and sharing one’s data. In the same post, I discovered the Open Science Framework which looks like an interesting website built on top of git.

Linklist October 1st 2012

  • If you know me, you know I do not like patents, as the recent cases in the tech industry exemplify. More clever people make the case to abolish them.
  • I do not know if I am a good programmer, but I like programming and would like to keep implementing some of my ideas myself, as Daniel Lemire argues in Why I still program (but this is even more controversial in fields such as transportation where programming is “just a tool”)

The Driverless Revolution

I had the opportunity for the Annual Workshop of the research chair on Mobility of my colleague Catherine Morency to make a short presentation of a favorite and slightly obsessive topic of mine, driverless cars. As you must know, they are coming and everyone should prepare for it. The slides are online (thanks to the Transportationist for his many posts).

Anything Gets Published

After first-hand experience of the randomness of the peer-review process, Daniel Lemire provided evidence that the process is basically useless. It does not guarantee that some quality standards are enforced or that original important papers will be published, with a huge inertia to accept really groundbreaking work. It is therefore a huge waste of everyone’s time which suggests that we would not be so much worse off if moving to a publish-then-filter system, à la

But I had never realized how low the quality of published work could go before I received an application by a student, with a list of publications including papers presented at conferences and published in lectures notes by Springer. One is entitled “Necessity of Accurately Registration Parameters Calculation for Good Reconstruction in Super‐resolution from a set of low images”, another contains the term “impaction” which is probably not what the authors meant for an image processing paper… Obviously the authors are not native English speakers, but the fact that no one highlighted the problem nor made sure that it was corrected is appalling. This is just one more glaring example that peer-review does not work.