Face recognition and detection plays an important role in biometric security in the modern era. It is also a very trending topic in the area of computer vision. Man has been trying to achieve more and more accuracy since the very beginning of image processing. But for a low end embedded application the task has always been quite challenging. Without a suitable CPU and GPU the task becomes very hard. Moreover, if the amount of data for training is also less then the task becomes quite impossible. Our thesis work aims to recognize and detect faces with a minimal amount of data for a low end embedded system application.