AI's Facial Expressions: A Mixed Bag of Success with Imagen-v2
- 10 minutes read - 2003 wordsTable of Contents
Facial expressions are a powerful tool for conveying emotions and intentions. In the realm of generative AI, the ability to create realistic and expressive faces is a crucial step towards creating truly immersive and engaging experiences. This blog post explores the capabilities of a generative AI model in generating facial expressions, analyzing its performance across various scenes and camera positions. We’ll delve into the model’s strengths and weaknesses, highlighting its successes in understanding scene context and aesthetics, while also uncovering its limitations in accurately capturing camera positions. Join us as we explore the exciting potential and ongoing challenges of AI in the realm of facial expressions.
Created with: imagen-v2
Intrigued by the Lights: A Portrait of Mystery
A close-up portrait captures a young woman with blonde hair, her gaze directed to the left, her expression a blend of intrigue and playfulness. The blurred background of a vibrant carnival or fair, with its colorful lights, adds a sense of mystery and movement to the scene. The dramatic lighting enhances the mood, leaving the viewer wondering what secrets lie within her enigmatic gaze.
Prompt
facial-expressions Amusement: Playful, carefree ; A lone woman; eye-level; Single Person; a bustling carnival with bright lights and colorful tents; cinematic
Characteristic
Shot : Close-up portrait of a woman with blonde hair, looking at the camera with a surprised expression. The background is blurred, with bright, colourful lights.
Aesthetic Score : 0.7
Mood : intrigued, mysterious, playful
Quality
Entropy : 6.66
Noise : 64
Prompt Clip Score : 0.27
AI Evaluation
Likelihood of AI : 0.80
Image errors : The image seems to be overly processed and sharpened, which makes some of the details appear artificial, especially in the hair and skin. The subject’s eyes appear a bit too large and unnatural. The overall image suffers from slightly unnatural, overly bright highlights.
Superman’s Jaw Drops at the Amusement Park
A costumed Superman stands in awe, his wide-open mouth and intense gaze capturing the excitement of the amusement park. The shallow depth of field draws attention to his dramatic reaction, while the blurred background hints at the thrilling rides and colorful atmosphere.
Prompt
facial-expressions Amusement: Exuberant, triumphant ; A superhero in a vibrant costume; eye-level; Hero; a crowded amusement park with roller coasters and Ferris wheels in the background; cinematic
Characteristic
Shot : A man dressed as Superman stands in a carnival, looking up and screaming, with a roller coaster in the background.
Aesthetic Score : 0.6
Mood : dramatic, intense, anxious
Quality
Entropy : 6.78
Noise : 73
Prompt Clip Score : 0.30
AI Evaluation
Likelihood of AI : 0.70
Image errors : The image has some artifacts around the edges of the man’s face and the roller coaster. Some parts of the image appear blurred, which might be a deliberate stylistic choice or a result of poor image quality.
Friendship, Laughter, and a Carousel: Capturing the Essence of Youth
Three friends bask in the warm glow of a summer afternoon, sharing laughter and stories on a blanket in the park. The carousel in the background adds a touch of whimsy, while the low-key lighting creates an intimate and playful atmosphere. This image captures the essence of youthful camaraderie and carefree joy.
Prompt
facial-expressions Amusement: Relaxed, happy ; A group of friends; eye-level; Normal People; a picnic blanket under a shady tree in a park, with a carousel in the distance; cinematic
Characteristic
Shot : Three young adults are laying on a blanket in a park, a carousel is in the background. The scene is shot from a low angle, giving a close-up perspective of the people.
Aesthetic Score : 0.6
Mood : casual, relaxed, playful
Quality
Entropy : 6.64
Noise : 100
Prompt Clip Score : 0.20
AI Evaluation
Likelihood of AI : 0.20
Image errors : There are no significant artifacts or errors in the image. The quality is good. The editing choices like the vintage filter might not be appealing to all.
Lost in the Game: Neon Lights Illuminate a Gamer’s Focus
A young man, headphones on, eyes glued to the screen, embodies the intensity of gaming. Neon lights cast a dramatic glow, highlighting his determined expression as he navigates the virtual world.
Prompt
facial-expressions Amusement: Focused, excited ; A gamer; close-up; Gamer; a dimly lit room with a computer screen displaying a vibrant video game, a controller in their hand; cinematic
Characteristic
Shot : A young man wearing headphones, possibly playing a video game, with a focused expression
Aesthetic Score : 0.6
Mood : intense, focused, determined
Quality
Entropy : 6.24
Noise : 80
Prompt Clip Score : 0.22
AI Evaluation
Likelihood of AI : 0.30
Image errors : The image appears to have some artificial sharpening and a slight noise reduction applied. The colors are slightly oversaturated.
Caught in the Moment: A Startled Glance
A young woman’s wide eyes and open mouth capture a moment of surprise, her gaze fixed on the camera. The blurred carousel behind her adds to the sense of uncertainty, leaving the viewer wondering what has caught her attention.
Prompt
facial-expressions Amusement: Magical, Thrilling ; A person; eye-level; Single Person; a carousel with brightly painted horses, her eyes wide with wonder; cinematic
Characteristic
Shot : A young woman with long blonde hair looks startled, her eyes wide with fear. The background is blurry and out of focus, but appears to be a carousel.
Aesthetic Score : 0.4
Mood : fear, suspense, anxiety
Quality
Entropy : 6.60
Noise : 69
Prompt Clip Score : 0.25
AI Evaluation
Likelihood of AI : 0.80
Image errors : The image suffers from a slight amount of blurriness, particularly around the edges and in the background. This may be intentional for artistic effect, but it does detract from the overall sharpness of the image.
Laughter and Fresh Produce: A Vibrant Marketplace Moment
Three friends share a joyful laugh amidst the colorful chaos of a bustling marketplace, surrounded by fresh produce and vibrant energy. The scene captures the essence of friendship, happiness, and the simple pleasures of life.
Prompt
facial-expressions Amusement: Joyful, carefree ; A group of friends, laughing and enjoying a sunny afternoon at a bustling outdoor market, surrounded by colorful stalls and the aroma of fresh food.; cinematic
Characteristic
Shot : Three people laughing in a market setting with lots of produce in the foreground, brightly colored umbrellas behind them
Aesthetic Score : 0.6
Mood : happy, joyful, vibrant
Quality
Entropy : 6.58
Noise : 83
Prompt Clip Score : 0.26
AI Evaluation
Likelihood of AI : 0.90
Image errors : The image has a slight painterly effect, which is noticeable in the skin, hair, and the produce, especially the tomatoes. There is a slight blurriness on some of the edges of the image. The color saturation is a bit high, giving an unnatural feel to the skin tones and the produce.
Lost in Thought: A Man’s Pensive Gaze
A solitary figure stands on a beach, his serious expression and the blurry ocean behind him creating a sense of melancholy and intrigue. The dramatic lighting and the man’s pensive mood evoke a sense of mystery, leaving the viewer wondering what thoughts are swirling in his mind.
Prompt
facial-expressions Amusement: Melancholy, contemplative ; A lone man; eye-level; Single Person; a deserted boardwalk at night, the sound of crashing waves in the background; cinematic
Characteristic
Shot : A man is standing outdoors, likely by the ocean, with the sun setting in the background.
Aesthetic Score : 0.7
Mood : mysterious, pensive, brooding
Quality
Entropy : 6.85
Noise : 115
Prompt Clip Score : 0.21
AI Evaluation
Likelihood of AI : 0.20
Image errors : There are some minor imperfections in the image, such as a slight blurriness around the edges of the man’s face.
Heroic Stance Amidst Chaos
A powerful superhero stands tall against a backdrop of fiery destruction, their unwavering resolve radiating in the face of danger. The dramatic explosion adds a sense of urgency and power to the scene, highlighting the hero’s strength and courage.
Prompt
facial-expressions Amusement: Thrilling, heroic ; A superhero in action; dynamic shot; Hero; a cityscape with towering buildings, a dramatic explosion in the background; cinematic
Characteristic
Shot : A muscular superhero stands in front of a cityscape. The city is on fire, with smoke and explosions in the background. The superhero is looking determined and ready to fight.
Aesthetic Score : 0.7
Mood : epic, heroic, dramatic
Quality
Entropy : 6.68
Noise : 64
Prompt Clip Score : 0.21
AI Evaluation
Likelihood of AI : 0.80
Image errors : The image has a few artifacts, particularly in the textures of the superhero’s costume, and also in the background cityscape. The edges of the subject are a bit fuzzy, possibly from oversharpening.
Gazing into the Unknown: A Moment of Shared Wonder
Three young friends stand silhouetted against a hazy sky, their faces turned upwards in a shared expression of curiosity and anticipation. The cropped composition draws the viewer’s eye to their gaze, leaving the source of their wonder a tantalizing mystery. This image evokes a sense of hope and excitement, capturing the essence of youthful exploration.
Prompt
facial-expressions Amusement: Exhilarating, bonding ; A group of friends, eye-level, enjoying a vibrant street festival, their faces lit up with excitement as they watch a lively performance.; cinematic
Characteristic
Shot : Three young people, possibly a group of friends, look up in awe or excitement, possibly at a performance or event. The background is blurry and indistinct, but there is a hint of a blue sky.
Aesthetic Score : 0.7
Mood : excited, curious, joyful
Quality
Entropy : 6.68
Noise : 103
Prompt Clip Score : 0.24
AI Evaluation
Likelihood of AI : 0.20
Image errors : The image has some minor artifacts, particularly in the hair and around the edges of the subjects. These are likely due to image compression or editing.
Caught in the Heat of the Moment: A Scream of Frustration
A close-up shot captures a man’s raw emotion as he screams into his headphones, his face contorted with frustration. The dark, blurred background adds to the intensity of the moment, leaving the viewer to wonder what drove him to this point.
Prompt
facial-expressions Amusement: Triumphant, exhilarating ; A gamer; close-up; Gamer; a dimly lit room, their hands moving rapidly on a keyboard, a triumphant shout escaping their lips; cinematic
Characteristic
Shot : A man is wearing headphones and screaming, his face is contorted in anger. There is a blue light reflecting on his face.
Aesthetic Score : 0.4
Mood : intense, angry, frustrated
Quality
Entropy : 6.15
Noise : 101
Prompt Clip Score : 0.28
AI Evaluation
Likelihood of AI : 0.10
Image errors : The image has some noise, particularly around the edges and in the shadow areas, and some unnatural sharpening.
Conclusion
The results show that the generative AI model performed well in understanding the scene and camera position, but struggled with the aesthetic aspect. Here’s a breakdown:
- Camera Position: The model scored 0.25, which is considered below average. This suggests that the model didn’t accurately capture the intended camera position described in the prompt.
- Shot Analysis: The model scored 0.55, which is considered good. This indicates that the model was able to understand the scene described in the prompt and create a shot that aligns with it.
- Aesthetic Analysis: The model scored 0.125, which is considered very good. This means that the generated image closely matched the expected aesthetic style.
Overall, the model demonstrated a good understanding of the scene and its aesthetic, but struggled with accurately capturing the intended camera position.
Sources:
- https://dramaresource.com/storytelling/
- https://seedsoftellers.eu/resources/the-body-language-for-young-tellers/
- https://digitalcollections.sit.edu/cgi/viewcontent.cgi?article=1288&context=sandanona&filename=1&type=additional
- https://citeseerx.ist.psu.edu/document?doi=7f842882e9bb1fa2c0e96939bc8d2c37e34e17c0&repid=rep1&type=pdf
- https://www.twinkl.co.uk/search?q=drama+facial+expression
- https://deepmind.google/technologies/imagen-2/