Towards a super-resolution based approach for improved face recognition in low resolution environment