Getting openSMILE to run on Android was a bit of a pickle. This rambling post is not intended as a step-by-step how-to, but is more intended as a collection of notes to help others. In the end I was able to run SMILExtract from openSMILE-2.1.0 on Android without using an NDK based Android project on a non-rooted phone. I have not attempted any feature extraction yet on Android, so it might still fail. A list of some of the software I used:

  • ndk-r10e
  • Phone running Android 4.1 (API level 16)
  • Android Studio v1.4.0
  • Ubuntu 14.04

Read More

As part of my research I had a look at the simulation of particles using the Discrete Particle Method (DPM), or as it is referred to more often: Discrete Element Method (DEM). The concept behind this simulation is based on calculating all forces acting on each particle at discrete time steps. I have described two simulations here; a Galton board and a collapsing cube of spherical particles. These simulation were just for fun and do not represent and real physical systems.

Collapsing cube

During one of my internet browsing sessions I came across the simulation of a regular cubic assembly of spherical particles which was allowed to collapse on a slight angular plane. I thought it looked rather awesome and decided to mimic it as best as possible.

Read More

I had an external hard drive lying around which I hadn't used for a couple of years. The drive was encrypted and I had no idea what my passphrase was anymore so the data was lost. I decided to avoid this in the future and had a look at how I can access my data if I forget my passphrase or the LUKS header gets corrupted.

When you know the master key it is possible to access your encrypted data even if you forgot your passphrase or the LUKS header is corrupted. In the remainder of the post I will show you how this is done. First of all a 10 MiB image is created for doing the experiments.

Read More

I wanted to create a polaroid effect using ImageMagick and discovered there is a -polaroid transformation operator to do this. Though the result looks nice, it was not polaroidish enough for me. A real polaroid picture has to have a larger border at the bottom to my opinion. I found the original work which led to the polaroid transformation operator. I used this and tweaked it a little to get a larger border at the bottom.

Custom polaroid effect Out-of-the-box effect
Custom polaroid effect on the left and out-of-the-box polaroid effect on the right.

The parameters used to create the images shown above is given below.

# Custom polaroid effect
convert -size 334x406 canvas:"#f0f0ff" -background none \
lena.png -geometry 300x300+17+17 \
-composite -interpolate nearest \
-rotate 90 -wave 5x812 -rotate -90 \
\( +clone -flop -background grey75 -shadow 70x5+15+15 \) \
+swap -background none \
-layers merge -resize 144 -rotate 5 polaroid1.png

# Out-of-the-box polaroid effect
convert lena.png -thumbnail 130x130 \
-bordercolor white -background grey75  -polaroid 5 polaroid2.png

For my final masters thesis I am currently working on finger vein authentication. Vein authentication is an upcoming type of biometrics which uses the vascular pattern of a certain body part as an unique unique property. As with all biometrics the idea is that no two person will exhibit the same vascular pattern. Currently the vascular patterns of the palm and finger are used for biometrics. Advantages of using vascular patterns for biometrics is that they are difficult to forge. For example a vascular pattern is not left behind after touching a surface as is the case with fingerprints. Another advantage the (probable) higher accuracy rate in terms of false rejections and false acceptances.

Vascular pattern of the finger
Vascular pattern of the finger
Read More