Conclusion

 

Our experiment yielded promising results, and fulfilled most of our goals set in the proposal.

 

We were able to code up an empirical version of the Shah snake, and provide an initial point selector for the snake.

 

We were able to improve upon this code but introducing a blur function and adding an additional coefficient.

 

We were able to qualitatively show the differences between different feature detection algorithms, and propose a method for developing quantitative results. 

 

Future studies could be done on:

            Dynamic insertion of snake nodes

            Automated parameter selection

            Established model of quantitative comparison