How does what we see affect our emotions?
Can a computer tell at a glance the difference between a cheerful image and a depressing image? Can he distinguish a romantic comedy from a horror movie in a matter of milliseconds? The question was answered by research published by University of Colorado Boulder neuroscientists: A computer, like our brain, can do this.
Tor Wager, one of the authors of the article, said that machine technologies are increasingly successful in recognizing the content of images and understanding what kind of object the image is. Stating that they are interested in whether a machine can do the same thing through emotions, Wager said the results of the researches were positive.
The results of the work on the innovation of machines and studies on imaging the human brain were published as an article in the journal Science Advances. It was recorded that the study on emotions using computer systems similar to the human brain would open the way for future research on the subject. The research also sheds a new and different light on how and where images are depicted in the human brain: Even what we see for a very short time is thought to have a greater and faster effect on our emotions than we can imagine. “Many people assume that people evaluate their environment in a certain way and that emotions come from old brain systems, such as the limbic system, Phil said Philip Kragel, co-ordinator of the study.
The emergence of EmoNet
For this work, Kragel began working with an existing neural network called AlexNet, which allows computers to recognize objects. Using previous research that describes the stereotypical emotional responses to images, he reconstructed the network to predict what a person would feel when he saw a particular image. Later, the new network, EmoNet, showed 25,000 images, ranging from sexual photos to nature landscapes, and asked them to classify them into 20 emotion categories such as passion, sexual desire, fear, and astonishment.
EmoNet was able to classify 11 of the emotion types accurately and consistently. However, he was better at recognizing some than others. For example, he found photos that aroused passion or sexual desire with an accuracy of more than 95 percent, but he was a little more challenged with more sensitive emotions. EmoNet was able to predict a feeling through colors: When EmoNet saw a black screen, it showed concern. Red evokes passion, purple fun. He chose romance if there was both red and purple. EmoNet was able to evaluate images correctly in general. When researchers showed EmoNet sections of short films and asked them to classify them as romantic comedy, action or horror movies, EmoNet categorized three quarters of these images correctly.
Images from functional magnetic resonance imaging (fMRI) machine
What you see is also what you feel!
The researchers also experimented with 18 people to further test and develop EmoNet. For a functional magnetic resonance imaging (fMRI) machine to measure brain activity, subjects were shown 112 images for four seconds. EmoNet was shown the same photos. When the activity in the neural network was compared with the brain of the subjects, the models were found to overlap. Ra We found a similarity between the models of brain activity in the occipital lobe and the units in EmoNet that encode specific emotions. That means EmoNet has learned to express emotions in a biologically appropriate way, even though we haven't taught it explicitly..
Brain imaging itself has produced some surprising findings. Even a short, basic image activated the emotion-related region in the visual cortex of the brain. It was observed that different emotions affected different regions during the study. “This shows that emotions aren't later add-ons in different parts of the brain, W Wager said.
Finally, researchers say that neural networks such as EmoNet can be used to help people digitally distinguish negative images or find positive ones. According to this; EmoNet can also be used to improve computer-human interactions and to help advance emotion-related research.
35AXX
Carex Day-Light Sky Bright Light Therapy Lamp - 10,000 LUX - Sun Lamp To Combat Winter Blues and To Increase Your Energy
