Abstract— Manual Facial Action Coding studies (FACS) have discovered a fuzzy facial expression that is both specific and sensitive to pain. However, manual pain coding imposes limitations such as training time and effort, technological requirements and human subjective factors. To surmount these challenges, in the last decade and a half, devices embedded with artificial neural networks (ANNs) have been used in researching pain through facial expression. Using neural- network theory, this paper argues that face perception of pain is organized around ‘fuzzy’ cases such that human observers judge a pain face based on their recognition that one face is more or less similar to other faces whose results are remembered and assessed (‘fuzzy case based reasoning’). A study implementing a fuzzy case-based reasoning system integrated with an ANN (FCBR-ANN) produced more than 90% accuracy in pain perception. Face perception of pain using an FCBR-ANN may be a real-time alternative to manual coding of pain by human observers, and may prove clinically useful. Index Terms— artificial neural network, CBR, fuzzy case- based reasoning, pain detection I. INTRODUCTION ACIAL expression is a major means for human beings to express emotions. The face can express emotion sooner than people verbalize or even realize their feelings. In the past decade, much progress has been made in building computer systems to understand pain through facial expression. However, much less is known about pain compared with emotional expression. Several studies using the Facial Action Coding System (FACS) have reliably identified the occurrence of certain combinations of facial muscles, contractions, or facial action units (AUs), across various acute clinical pain conditions [1][2]. In general, AU's are a contraction or relaxation of one or more facial muscles. There is consensus regarding the claim that the facial expression of pain is distinct from the expression of basic emotions [3]. According to FACS investigator's guide, AU4 (brow lower), AU6 (cheek raiser), AU7 (lid tighten), AU9 (nose wrinkle) and AU10 (upper lip raiser) are the target action units that occur when pain is facially expressed. Moreover, the following AU12 (lip corner puller), AU20 (lip stretch), AU25 (lips part), AU26 (jaw drop) and AU27 (mouth stretch) may occur with pain and/or with major variants. The set of used AU’s within this work is presented in Table I. Manuscript received March 18, 2016. Mati Golani is with Ort Braude College, Department of Software Engineering, P.O. Box 78 Snunit 51, Karmiel 21982 Israel (phone: + 972- 4-9086464; fax: +972-4- 9901-852; e-mail: matig@braude.ac.il). Simon P. van Rysewyk is with University of Tasmania, Department of Philosophy, School of Humanities, (e-mail: simon.vanrysewyk@utas.edu.au, simon@rirl.org). TABLE I COMMON AU’S AU description Facial muscle Example 26 Jaw drop Masseter, relaxed Temporalis and internal Pterygoid 4 Brow lower Corrugator supercilii, Depressor supercilii 43 Eyes closed Relaxation of Levator palpebrae superioris; Orbicularis oculi, pars palpebralis 1 Inner brows raiser Frontalis, parsmedialis 15 Lip corner depressor Depressor angulioris (a.k.a. Triangularis) 20 Lip stretcher Risorius w/ platysma 9 Nose wrinkler Levator labii superioris alaquaenasi 2 Outer brow raiser Frontalis, pars lateralis 5 Upper lip-raiser Levator palpebrae superioris This paper argues that ANN approaches to face perception of pain is organized around ‘fuzzy’ cases such that human observers judge a pain face based on their recognition that one face is more or less similar to other faces whose results are remembered and assessed. It is clear that there is an increased demand for rapid and accurate pain detection in the age of remote medicine and mobile computing, where emerging technologies can be adopted in order to improve patient treatment and satisfaction. This paper is organized as follows: Section IIII introduces known structured and non-structured based modeling methods, including fuzzy CBR-ANN based model. In Sections III,IV,V we present data pre-processing and Pain Perception - a Fuzzy CBR Approach Mati Golani, Simon P. van Rysewyk F Proceedings of the World Congress on Engineering 2016 Vol I WCE 2016, June 29 - July 1, 2016, London, U.K. ISBN: 978-988-19253-0-5 ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online) WCE 2016