Detection of Human Face by Thermal Infrared Camera Using MPI model and Feature Extraction Method Chirag Kumar Kyal Computer Science and Engineering National Institute of Science and Technology Berhampur, India chirag.kyal@gmail.com Harsh Poddar Computer Science and Engineering National Institute of Science and Technology Berhampur, India harshpoddar1549@gmail.com Motahar Reza Computer Science and Engineering National Institute of Science and Technology Berhampur, India reza@nist.edu Abstract – Thermal imagery is a substitute of visible imagery for face detection due to its property of illumination invariance with the variation of facial appearances. This paper presents an effective method for human face detection in thermal imaging. The concept of histogram plot has been used in the feature extraction process and later in face detection. Techniques like thresholding, object boundary analysis, morphological operation etc. have been performed on the images to ease the process of detection. In order to enhance the performance of the algorithm and to reduce the computation time, parallelism has been achieved using Message Passing Interface (MPI) model. Overall, the proposed algorithm showed a higher level of accuracy and less complexity time of 0.11 seconds in the parallel environment as compared to 0.20 seconds in a serial environment. Keywords- Thermal Imaging, Face Detection, MPI model, Image Processing, Histogram between pixel intensities. I. INTRODUCTION In recent years, Face detection technology has gained an increasing attention due to its numerous applications related to video surveillance system, security information, military fields, fraud detection, and object detection (e.g. ability to distinguish an object and a human). Most of the research work on face detection are limited to visible images. Face detection in the visible image has tackled various problems still, there are some problems which have not been solved completely using visible images. Such as (1) Face detection in different lighting conditions especially in low intensity of light. (2) Discrimination between printed faces and real human faces [1] (3) Emotion detection and stress detection using thermal images [2]. So, we need to find a better approach to overcome these issues. The thermal infrared (IR) camera detects heat emitted by the objects rather than reflected light. It converts the temperature of objects into colors of gray which are darker or lighter than the background. Generally, the temperature of human face is constant and higher than its background. This solves the previously discussed problems because thermal camera is only sensitive to thermal IR band while insensitive to visible light. Thermal imaging is also called IR imaging as it uses an electromagnetic radiation emitted from the heat, reflected by an object. Based on wavelength, the IR bands can be divided into 4 types: Fig 1: Different Kinds of IR band: (i) Near-Infrared (NIR) (ii)Short-Wave infrared (SWIR) (iii) Mid-Wave infrared (MWIR) (iv) Long-Wave infrared (LWIR) In thermal image, the lighter areas indicate higher temperature and the level of details in the image decreases with the increase in wavelength. Though the images are taken in Long-Wave infrared (LWIR) band, contains less detailed feature, it is most appropriate for invariant lighting condition. The long-wave infrared imaging can operate even in absolute darkness by applying electromagnetic sub-band spectrum (8- 14 μm) and thus can obtain illumination invariant image. In this paper, we have proposed an algorithm for face detection by feature extraction using thermal infrared camera. To the best of our knowledge, there are very few publications available on this topic. Therefore, we think of working in this field to solve some essential problems. The organization of the paper is the following: Dataset acquisition and creation is described in Section 2.1, followed by the proposed algorithm and technique for face detection in Section 2.2. Section 3 consists the Pseudo Code for the serial implementation of the proposed algorithm. Experimental results are drawn in Section 4. Finally, Section 5 reports the conclusion of our work. II. THE PROPOSED METHOD In this section, an algorithm for face detection in thermal images has been described. This work has been developed by using Python programming language along with the use of OpenCV library. Parallelism in the algorithm is obtained using the MPI model binding for Python, that is, mpi4py package. 2018 4th International Conference on Computing Communication and Automation (ICCCA) 978-1-5386-6947-1/18/$31.00 ©2018 IEEE 1