Published Paper Details:

2-D and 3-D models can be used for face recognition. The effect of light due surrounding factors can be reduced using 3-D model. But the issue with 3-D model method is that it takes large amount of time for fitting. This makes it difficult to meet real-time requirements. The algorithms usually use 2-D model for recognizing a face. These 2-D images are used for everything as the image captured can be of any distance using appropriate sensors and any illumination scenario. Whereas for capturing an image in 3-D, imaging technologies that currently work are able to do so at very short distances and also suffer from the inability to replicate complex patterns unless they get extremely close to the subject. 3-D imaging is a new technology when it comes to image capturing as its original purpose was to scan and measure objects accurately for aeronautical and scientific applications. Current commercial devices have cameras that can work well with low exposures and bad lighting situations using a combination of 2-D and 3-D imaging techniques that helps in creating incredibly detailed images. Our aim in this paper is to improve low-light camera performance using 2-D face recognition algorithms for extracting facial features captured using older security footage devices with no additional sensors. This paper is aimed to work with video streams as well as images so that plenty of learning and testing data is available to the paper.