The stress has an end :)

My last post has been a while ago! That’s been the result of heavy brain work activated by assignments.

Indeed, lectures have been less but the work load in the labs has potentially increased.

In Machine Learning lecture, supervised learning was finished and revised. The professor started with unsupervised learning. We got introduced to some super interesting applications like speech recognition or heavily compressed transmission of data by pattern classification with the K-Means algorithm.
The lab parts of the unsupervised learning parts were due On Wednesday night. I finished report and code of linear regression (curve fitting), logistic regression (classification) and neural network (brain like construct for prediction).

Linear Regression

Linear Regression – Overfitting

Logistic Regression – Clustering with decision boundary

Neural Network – XOR

I managed to generated ball coordinates with MATLAB. Now I have a ball rotating in my application – very very fancy!
Within the lecture back and front clipping algorithms were discussed for 3D spaces. Interesting although quite complex. I might implement that in my application. It’s not compulsory though.

When it comes to Introduction to Computer Vision I didn’t spend any time recently with that subject. The other subjects kept me too busy and I wanted to have some spare time for me too I hope this is not bouncing back on me!
Here are some pictures from my last assignment:

Intersection Histogram of two frames

Color Histograms of one Video Frame

Image Convolution – Averaging Kernel and Edge Detection Kernel

There’s only one subject left to give you an update: Advanced Transform Methods. The professor explained the Gabor Transform – please ask me in a couple of weeks what it is!
I was working hard in the labs to understand the Fourier Transform and to write a report for the assignment. I think I got a clue now. I generated spectrograms and other fancy things in the lab:

15 Hz Sinus Function 100 Hz Sampling Comparison

Fourier Transform of 15 Hz Sinus Function with 100 Hz Sampling Comparison

Spectrogram of Audio file with three core frequencies – Window Size 4096

Now you might think apart from studies nothing else. Well you are almost right. I’ve been in the cinemas Sunday night to watch Interstellar! It’s a quite deep movie but it was very good entertainment. Many times the movie surprises with the story although the overall story is the most common in Hollywood movies: The planet needs to be saved. At certain points I wonder about who was first: chicken or egg. The physics sometimes aren’t very clear but that doesn’t harm the entertainment value.
And on Wednesday I have been to Hyde Park and Leicester Square for a walk. I loved the Christmas decoration and flashing banners along Oxford Street and Picadilly Circus!

Enjoy the pictures!!