Some of my machine learning projects!~
These are some of the machine learning projects that I did for a machine learning module in my university. The project started of easily by classifying images which contains a person making an Italian hand sign to gradually predicting the hand signs in a video.
The datasets could be found in Chalearn: http://chalearnlap.cvc.uab.es/dataset/13/description/
The assignments were submitted in Kaggle competing with my fellow classmates.
The first assignment restricted us to use only Linear models. Thus, I played around with logistic regression, SVM and random forests and feature extracting techniques for images like HOG and SIFT. The one that yielded me the best accuracy results were a combination of random forests and dense SIFT techniques.
The second assignments requiured us to use neural networks specifically CNN since we are talking about images here. I used tensorflow here where as in my first assignment I used sci-kit environment.
In the third and final assignment, we were given a video as a dataset instead of image and we were tasked to use RNN for classification. I used a two layer LSTM stacked upon CNN together with data normalization and augmentation to perfom the classification.
Overall, this is a very interesting project and I learned a lot and hope to explore more new algorithms and techniques in the field of machine learning!~