Figure 3 - from "Facial Emotional Expressions of Life-like Character Based on Text Classifier and Fuzzy Logic"

Emotion generation in an algebraic model developed by 36] defined a set of parameters motivators. These motivators consist of keywords (mapped to emotional facial expressions, for example "happy” to smile”), hierarchical context to calculate coherence, user profile which associated with in- hibitor/amplifier affecting facial display and biological needs for example, blinking and wetting the lips). Motivators with their respective weights then were feed into a multiplexer to display a final facial expression. In our system, instead manual input of keywords, machine learning approach is employed to process the text. Fig. 3. Facial expression of neutral and basic emotions: sadness, joy, anger, fear, disgust, surprise (from left to right) [27].

subdirectory_arrow_right

Figure source:

Fig. 1. Supervised Text Classification. Input is first preprocessed using language-dependent tools, such as tokenizing, stop words removal and (optionally) lemmatization. (a) During training, a feature extractor is used to convert each input value to a feature set. Pairs of feature sets and classes are fed into the machine learning algorithm to generate a classification model. (b) During prediction, the same feature extractor is used to convert unseen inputs to feature sets. These feature sets are then fed into the classification model, which generates predicted class. During cross-validation, these predicted classes will be matched against ground-truth classes.

cosine similarity value) training document as the class of a query document.

A well known model of emotions is the work of [22]. He uses basic emotions as building blocks for derived emotions; secondary emotions and even ternary emotions. All other emotions are mixed or derivative states; that is, they occur as combinations, mixtures, or compounds of the basic emotions, as depicted in Fig. 2. Fig. 2. Plutchik’s Wheel of Emotion [23] believed there exists a relationship between facial expression and emotional state. The proponents of the basic emotions view [24], [25], according to [26], assume that there is a small set of basic emotions that can be expressed distinctively from one another by facial expressions. For

+20+19+18

instance, when people are angry they frown and when they are happy they smile. The six basic emotions defined by [24] can be associated with a set of facial expressions. [27] has designed a face model with 6 (six) facial expressions of basic emotions depicted in Fig. 3. Table II shows textual descriptions of facial expressions as representations of basic emotions, taken from [28]. In this paper, surprise is excluded due to unavail- ability of surprise class in the training set, however facial expression of shame is added based on following arguments from psychological researches.

+20+19+18

surya sumpeno

Facial Emotional Expressions of Life-like Character Based on Text Classifier and Fuzzy Logic

Abstract: A system consists of a text classifier and Fuzzy Inference System (FIS) to build a life-like virtual character capable of expressing emotion from a text input is proposed. The system classifies emotional content of sentence(s) from text input and expresses corresponding emotion by a facial expression. Text input is classified using the text classifier, while facial expression of the life-like character are controlled by FIS utilizing results from the text classifier. A number of text classifier methods are employed and their performances are evaluated using Leave-One-Out Cross Validation.

descriptionSee full PDF