About me

avatar1024Welcome. This is my portfolio site showcasing my scientific work, other projects and photography. I’m originally from The Netherlands but currently living and working as a PhD student at Justus-Liebig-University Giessen in Germany. Here I study the perception of liquids and other deformable materials using psychophysical techniques. Next to work I also have great interest in photography. Mostly while traveling I find time to pick up my camera as you might see in the photography section.

I have a background in computer science and a bit of computer/electrical engineering. After following mostly HCI (human computer interaction) oriented courses I ended up in the field of psychophysics. However after my PhD research I will try to steer my career back to more HCI oriented work where I can combine my knowledge in psychophysics and engineering. In the end I’m really interested in creating digital extensions of our physical world enabling experiences beyond the capabilities of our regular senses.

If you have more specific questions please don’t hesitate to contact me at mail [at] janjaap [dot] info.

Research: Liquids

Fluids and other deformable materials have highly mutable shapes, which are influenced by both intrinsic properties (e.g. viscosity, velocity) and extrinsic forces (e.g. gravity, object interactions). During my PhD in Professor Roland Fleming’s lab at Gießen University I studied the perception of liquids and how we are able to estimate their properties consistently across a wide range of contexts.


In my research I try to map the perceptual space and cues we use to perceive liquids. We find that especially mid-level shape/motion cues (e.g. spiralling, ripples, spread) are very informative of liquid properties and allow us to extract context invariant information. It is hard to physically measure concepts such as blobbiness or folding in a way that is scale and orientation invariant across 3D shapes. Therefore, we use neural networks to perceive viscosity in images and especially focus on networks that make similar estimation errors as human observers. By studying the learned properties of the networks we hope to gain insights on new image-based cues we use to achieve viscosity constancy. The larger goal is to achieve better generalisation in computer vision.

The Stimuli

I invested much time to set up a technical pipeline to generate stable, precise and realistic liquid stimuli. I used a particle simulation program RealFlow made for the VFX industry to simulate liquids. The computational costs of these stimuli are quite high. I wrote specific scripts to distribute the calculations over various systems and clusters. Over 4 million images were generated showing liquids with different viscosities and optical appearances interacting in different scenes. In 2018 I became a certified RealFlow Xpert.

Research: Gloss

For my Master studies at Leiden University in The Netherlands I did a graduation project on gloss perception. The actual research was done in the Perceptual Intelligence Lab at Delft University of Technology where Sylvia Pont and Maarten Wijntjes were my external supervisors. This was the first time I came in contact with psychophysical research.


We investigated the influence of the spatial structure of the illumination on gloss perception. The inspiration came from various art works like the paintings of Vermeer, where much simpler highlight shapes are used compared to a real world situations. We performed three experiments of which two took place using digital photographs on a computer monitor and one with the real spheres in a light box. The observers had to compare and rate glossiness. The results show that more complex highlight shapes were perceived to produce a less glossy appearance than simple highlight shapes such as a disk or square. These results show that, contradictory to some beliefs, the complexity of a highlight shape’s spatial structure alone is not the main criterion for an increase perceived glossiness.

The Stimuli

A diffuse light box in combination with differently shaped masks was used to produce a set of 6 simple and more complex highlight shapes. In the box we placed spherical stimuli that were painted in 6 degrees of glossiness. This resulted in a stimulus set of 6 highlight shapes and 6 gloss levels, a total of 36 stimuli. Observers were asked to rate glossiness looking at the real scene in the light box, but we also performed experiments with photographs of the stimuli. The figure below shows a subset of the stimuli. Six masks where chosen to represent common shapes used in studio photography, paintings and cartoons, like the disk, square, window and ring shape. The other shapes where chosen to see if highlight shape complexity increases the perceived glossiness.

stimuli selectie



2018 Jan Jaap R. van Assen, Pascal Barla, Roland W. Fleming (2018). Visual Features in the Perception of Liquids. Current Biology, 28(3), 452-458, doi: 10.1016/j.cub.2017.12.037. [PDF]

2017 Filipp Schmidt, Vivian C. Paulun, Jan Jaap R. van Assen, Roland W. Fleming (2017). Inferring the stiffness of unfamiliar objects from optical, shape, and motion cues. Journal of Vision, 17(3):18, 1–17, doi: 10.1167/17.3.18. [PDF]

2017 Vivian C. Paulun, Filipp Schmidt, Jan Jaap R. van Assen, Roland W. Fleming (2017). Shape, motion, and optical cues to stiffness of elastic objects. Journal of Vision, 17(1):20, 1–22, doi: 10.1167/17.1.20. [PDF]

2016 Jan Jaap R. van Assen, Roland W. Fleming (2016). Influence of optical material properties on the perception of liquids. Journal of Vision, 16(15):12, 1–20, doi: 10.1167/16.15.12. [PDF]

2016 Jan Jaap R. van Assen, Maarten W. A. Wijntjes, & Sylvia C. Pont (2016). Highlight shapes and perception of gloss for real and photographed objects. Journal of Vision, 16(6):6, 1–14, doi: 10.1167/16.6.6. [PDF]


2017 Jan Jaap R. van Assen, Pascal Barla and Roland W. Fleming. Shape features underlying the perception of liquids. Computational and Mathematical Models in Vision (MODVIS) Annual Meeting 2017.

2016 Jan Jaap R. van Assen, Pascal Barla and Roland W. Fleming. Identifying shape features underlying liquid perception. European Conference on Visual Perception (ECVP) Annual Meeting 2016.

2016 Jan Jaap R. van Assen, Pascal Barla and Roland W. Fleming. Cues Underlying Liquid Constancy. Visual Sciences Society (VSS) Annual Meeting 2016. Abstract|Slides

2015 Jan Jaap R. van Assen and Roland W. Fleming. The influence of optical material appearance on the perception of liquids and their properties. Visual Sciences Society (VSS) Annual Meeting 2015.
Abstract|Slides (22mb).


2015 Elsevier/Vision Research Travel Award

2013 ECVP 2013 Travel Grant Award

Other Projects

A selection of projects I worked on during my Bachelor and Master studies.

Light Architecture

One elective course I followed was Light Architecture at Delft University of Technology. We worked with seven students on this project with different backgrounds in industrial design and architecture. We had to design a light plan emphasising the architecture of in our case the H. Antonius Abt in Scheveningen. This church was build in 1927 with a very sober exterior and very rich interior. This is something we wanted to translate into the light design. The church was interested in a free detailed light plan and therefore provided building schematics and were part of the conceptual process. After the requirements were defined we started to make a light plan.

To express the design we used 3D renderings. Google Sketchup was used to make the model using the building schematics and the Vray render engine was used to render the light. It was at that time the newest version of Vray supporting IES profiles. An IES profile captures the specific light characteristics of an illuminant which is measured under lab conditions. This is used to simulate very realistic representations of an existing light source. We used lights made by ACDC lighting and implemented their IES files in our design. The rendering represented the light emitted of their existing products and therefore give a very good impression of the light design.


I was always interested in ferrorfluids and during my Master studies I got a chance to experiment with this during the course Hardware & Physical Computing. We didn’t had much resources to our disposal like a good electro magnet which required some creative solutions. A circuit was designed using TIP122 transistors, 3-Amp diodes and a modified transformer (cut in half) combined with permanent magnet.

The power source was a modified PSU (computer power supply) that delivered 6 Amps at 12 volts. Two LDRs were used as sensors to control the ferrofluid. The entire circuit was controlled by an Arduino where one LDR was controlling the amplitude of the magnetic field and one LDR the frequency of the pulse width modulated sinusoid. This modulation made it look like as if the ferrofluid was breathing.


This crazy piece of machinery was build to let you perceive the visual world around you in a more dynamic way making changes to the visual field. The eyes are being replaced by two motorized webcams which send their own signal to one of the eyes. These motorised effects are really strange to perceive since the ‘new’ eyes can turn to much more extremes than a normal eye. There is a small 10″ inch netbook build in the helmet. The display of the notebook is used to show the two webcam feeds. There is an Arduino on top with three servo motors. To keep the helmet wireless we need power.

There is a 9v battery for the Arduino and four heavy ‘D’ batteries to power the servo’s. We used Max/MSP Jitter to display the webcam feeds on the screen. We used the accelerometer of an Android phone to control the servo’s. The data is send trough wifi network locally set up on the laptop. The OSC protocol is used to send the data from the accelerometer to Max/MSP. The main idea was that the phone can also be fixed on the helmet and therefore your head movement will also control your eyes. The result, The Desyncapitator, is shown in the following video:

T3 Touch Table

T3 or Tastbare Taal Tafel (Tangible Language Table) is an interactive touch table with finger and object recognition. The table was designed for primary schools where children could write/tell/draw stories on it. This was part of the HCI (Human Computer Interaction) semester during my Mediatechnology study at the University of Applied Sciences of Utrecht. Reactivision was used for fiducial recognition and Community Core Vision for multitouch finger recognition. It is a multitouch table with object recognition and a flash application for the story telling.

There are some additional tools which are physical objects which can be placed on or removed from the screen. The diameter of the screen in somewhat limited because the height of the table. The users are children between 6 and 9 years old so the table can’t be that high. There is a projector placed inside the table and a mirror to extend the projection length and therefore the screen size. There is a webcam with infrared pass-trough filter and 6 infrared LED spots for the finger and object recognition. The video below shows the design and functions (in Dutch).

Footsteps of Mutation

Are the possibilities of delivering a meaningful sentence endless? Can sentence meaning transform into a higher level of poetic art through language? Experience the mutations of sentence meaning as Dutch sentences take their first footsteps to mutation. We needed to build an installation that depicted the main theme ‘obsolete’ and sub-theme ‘mutation’. We chose to put this in the context of language where the mutation of language results in meaning becoming obsolete. Our installation showed this by projecting mutated sentences on a canvas. For the mutation of language we used the Delilah system from Leiden University Centre for Linguistics.

This system can parse and generate Dutch semantically. So the user enters a sentence which then will be parsed by Delilah and from this parsed meaning a new sentence will be generated and projected on the screen. By doing this, sentences mutate and if they mutate too much their meaning becomes obsolete. We used an old typewriter to enter the sentences because typewriters became obsolete too with digital devices taking over. Some technical hurdles had to be taken to integrate a 1930’s typewriter, Delilah using SICStus Prolog and display the sentences on a canvas using processing and java. The final installation was very promising even though it could use bit more refinement.

Dot Matrix Organ

The assignment for the New Media & New Technologies course was to make a tribute to an old new media carrier. We (classmate Erik and me) chose the matrix printer, which was in its time very important because in combination with a pc it completely replaced the typewriter from the average home. We tried to connect various stepper motors from different printers to an arduino. This was some work since stepper motors work only by sending synchronized signals to the different magnets in the motor. Luckily there are some useful libraries availible. The plan was to connect the printer to a keyboard and generate different tone heights by pressing different keys.

My parents had a matrix printer from 20 years ago somewhere hidden on the attic we could use for the final product. The experience gained exploring other parts got useful connecting the Epson. The stepper motors used in the Epson turned out to be more advanced with a bipolar instead of unipolar design. We had to change some circuit plans including the ULN2003 chip which increases the current over line using a Darlington array. After figuring out the right circuit we only needed to fine tune the software part and connect the keyboard which was much easier. The result although not musically fine-tuned worked perfectly.

  • Light Architecture

    Light Architecture

  • Ferrofluids


  • Desyncapitator


  • T3 Touch Table

    T3 Touch Table

  • Footsteps of Mutation

    Footsteps of Mutation

  • Dot Matrix Organ

    Dot Matrix Organ


Below you will find a selection of my work. For more photos go to my 500px account here.