A. Realization of requirements: 1. use opencv with data training approach to learn samples of the letters in order for opencv to recognise 2. problem space is red from dictionary (bundled with app, no integration with 3rd party) 3. astar or similar algorithm is used to find the best solution according to one criteria (no multicriteria optimisation) B. Project assumptions: 1. learning sets for letter recognition are delivered by the employer 2. accuracy of recognition is limited to capabilities of opencv and quality of learning sets delivered by employer 3. rough UI, no focus on candy and nice UI - focus on functionality not on look & feel 4. possible location of target letters is fixed, only one letter per image 5. i'm not responsible for potential letter recognition issues in opencv lib, i'll do my best to make it work for the project but i can't guarantee quality. It may need some time to finetune the opencv and it's not included in the offer 6. size of the problem space which is traversed by path algorithm is limited to capabilities of iPhone memory/cpu. Too large space may cause performance issues and slow recognition. 7. 3 milestones in project: opencv lib ready for learning sets, letters recognition, path finding and solution presentation