Emotion Recognition Using Emotiv Sensorntg.ailab.wsu.edu/class_posters/2010/EEG_Emotions.pdfThe...
1
INTRODUCTION Our emotional state plays a key role in how we experience and interact with the environment. Emotion directly affects our decision making, perception, cognition, creativity, attention, reasoning and memory. The estimation of emotional changes from EEG is of great interest among researchers and people who develop devices in the HCI field. We conducted experiments using the Emotiv Headset to find variations in different emotions while participant watches different movies. We extracted the raw data from the Headset. The comparative analysis reveals that the raw data is helpful in revealing patterns of emotions which we will use in our second phase. DATA COLLECTION The participant was a normal 23 year old, right handed without any brain trauma. The data was collected from 14 sensors for approximately 12 minutes. Each tests was conducted consecutively for approx. 120 seconds. Five such experiments were conducted. Emotion Recognition Using Emotiv Sensor Raghavendiran Srinivasan and Thomas Corll School of Electrical Engineering and Computer Science Faculty: Diane J. Cook and Maureen Schmitter-Edgecombe FUTURE PLANS The headset shows promising results of giving emotional data. We are still not convinced of the headset’s ability. To help ensure success of the project, we will add Gyroscope data. Hopefully, the Gyroscope data will not be needed, but provide a safety net. We will use the open source program PokerTH as the Poker Engine. To store poker and EEG data we will use SQLite. Two human players will use the Emotiv headset. The experiment will have two phases training and testing. During training, the humans and an AI agent will play each other. Machine Learning will classify if a player is bluffing or not . In the test phase this information will give the agent an advantage. The Emotiv headset classifies the following emotions: excitement, engagement/boredom, meditation and frustration. The raw data of emotions (in percent) is collected. The emotional data is generated in real time in millisecond increments. Different genres of movie trailers were watched with the Emotiv headset on. Figure 3 shows the emotional data collected while watching five movie trailers. We observe that the participant was very excited while watching Source code and Megamind. And it was verified by the participant._______________________________ Figure3. Variation of Emotions Figure 1. EMOTIV Headset Figure 2. Sensor Raw Data (a) Motion Sensor *** EXPERIMENTS & RESULTS
Transcript of Emotion Recognition Using Emotiv Sensorntg.ailab.wsu.edu/class_posters/2010/EEG_Emotions.pdfThe...
-
INTRODUCTION
Our emotional state plays a key role in how weexperience and interact with the environment. Emotion directly affects our decision making,perception, cognition, creativity, attention, reasoningand memory. The estimation of emotional changes from EEG is ofgreat interest among researchers and people who developdevices in the HCI field. We conducted experiments using the Emotiv Headsetto find variations in different emotions while participantwatches different movies. We extracted the raw data from the Headset. The comparative analysis reveals that the raw data ishelpful in revealing patterns of emotions which we willuse in our second phase.
DATA COLLECTION
The participant was a normal 23 year old, righthanded without any brain trauma. The data was collected from 14 sensors forapproximately 12 minutes. Each tests was conducted consecutively for approx.120 seconds. Five such experiments were conducted.
Emotion Recognition Using Emotiv SensorRaghavendiran Srinivasan and Thomas Corll
School of Electrical Engineering and Computer ScienceFaculty: Diane J. Cook and Maureen Schmitter-Edgecombe
FUTURE PLANS
The headset shows promising results of givingemotional data. We are still not convinced of theheadset’s ability. To help ensure success of the project, we will addGyroscope data. Hopefully, the Gyroscope data will notbe needed, but provide a safety net. We will use the open source program PokerTH as thePoker Engine. To store poker and EEG data we will use SQLite. Two human players will use the Emotiv headset. Theexperiment will have two phases training and testing. During training, the humans and an AI agent willplay each other. Machine Learning will classify if a player is bluffing ornot . In the test phase this information will give theagent an advantage.
The Emotiv headset classifies the following emotions:excitement, engagement/boredom, meditation andfrustration. The raw data of emotions (in percent) is collected. The emotional data is generated in real time inmillisecond increments. Different genres of movie trailers were watched withthe Emotiv headset on. Figure 3 shows the emotional data collected whilewatching five movie trailers. We observe that the participant was very excitedwhile watching Source code and Megamind. And it wasverified by the participant._______________________________
Figure3. Variation of Emotions
Figure 1. EMOTIV Headset
Figure 2. Sensor Raw Data
(a) Motion Sensor ***
EXPERIMENTS & RESULTS
![Investigation of Emotiv Epoc+ Headset Usability by ...€¦ · Fig.3: Headset Emotiv Insight. [6] 2.3 Emotiv Epoc Flex Emotiv Epoc Flex is a wireless EEG headset. Possible sensors](https://static.fdocuments.net/doc/165x107/5f3a3e29970ae26f83602061/investigation-of-emotiv-epoc-headset-usability-by-fig3-headset-emotiv-insight.jpg)
