cd Eigenfaces
mkdir data/celeb
mkdir data/raw
mkdir data/output
python preprocessing/download.py
python preprocessing/celeb.py
rm -rf data/raw
```
## Run
```
./main
```
Contains the fist part of the assignment, computing the accuracy over a
60% training/ 40% testing partition of the data with at most 85% accuracy and removing the
first eigenvector.
```
./best_five
```
Takes 10 celebrity pictures and returns the five best matches for each,
storing the data in the file data/celeb_accuracy_scores, and displaying 5 of the celebrities
together with the closest matches on the screen
```
./celeb
```
A simplified version of ./best_five, only displaying the best match and without
storing the accuracies.
```
./varyEnergy
```
Runs 20 tests with varying energy, both including and excluding the first eigeface.
Additionally, the script in postprocessing creates the data using the output data generated with ./varyEnergy executable
mihaifieraru/Eigenfaces
This is an implementation of the eigenfaces algorithm (http://www.cs.ucsb.edu/~mturk/Papers/jcn.pdf)
C++