AI's Artistic Vision: Capturing the Scene, But Missing the Shot with Midjourney
- 9 minutes read - 1837 words
Table of Contents
In the realm of artificial intelligence, image generation has emerged as a captivating field, with models capable of creating stunning visuals. However, the accuracy of these models in capturing specific details, such as camera positioning, remains a subject of ongoing research. This blog post explores the results of a study that tested an AI model’s ability to generate images based on detailed prompts, focusing on its performance in capturing scene aesthetics and camera positioning. The findings reveal a fascinating interplay between the model’s strengths and weaknesses, highlighting the potential and challenges of AI in image generation.
Created with: midjourney
Silhouette of Hope in a Mysterious Hallway
A solitary figure stands silhouetted at the end of a long, shadowy hallway. A window at the far end casts a sliver of light, hinting at a world beyond the darkness. The scene evokes a sense of mystery, isolation, and perhaps, a glimmer of hope.
Prompt
key-lighting High-key lighting with a single, strong light source: Mysterious, introspective ; A lone figure standing in a doorway; medium-shot; Single Person; A dimly lit hallway with a single light source illuminating the doorway; cinematic
Characteristic
Shot : A woman stands silhouetted in a dark hallway, looking towards a bright window at the end of the hall.
Aesthetic Score : 0.6
Mood : mysterious, lonely, pensive
Quality
Entropy : 5.44
Noise : 99
Prompt Clip Score : 0.21
AI Evaluation
Likelihood of AI : 0.20
Image errors : No significant errors, but the image is slightly overexposed in the doorway, which may be intended.
Hero Stands Alone Against the Inferno
A lone superhero, silhouetted against a fiery cityscape, evokes a sense of dramatic tension and impending doom. The scene is both epic and somber, capturing the weight of the hero’s struggle against overwhelming odds.
Prompt
key-lighting Backlighting with a strong, warm light source: Epic, heroic, dramatic ; A superhero silhouetted against a blazing sunset; medium-shot; Hero; A cityscape with towering buildings and a fiery sky; cinematic
Characteristic
Shot : A silhouetted figure stands in front of a burning city, with a fiery sky above. The figure appears to be a superhero with a cape, facing the devastation.
Aesthetic Score : 0.6
Mood : dramatic, apocalyptic, heroic
Quality
Entropy : 6.73
Noise : 100
Prompt Clip Score : 0.29
AI Evaluation
Likelihood of AI : 0.80
Image errors : The image has some minor artifacts, particularly around the edges of the buildings and the fire.
A Moment of Solitude: Light and Shadow in the Kitchen
A young woman finds peace in the quiet of her kitchen, bathed in warm window light. The contrast between light and shadow creates a dramatic effect, highlighting her silhouette and the bowl of food she enjoys. The scene evokes a sense of calm, pensive solitude.
Prompt
key-lighting Soft, diffused lighting with warm tones: Peaceful, everyday, intimate ; A young woman sitting at a kitchen table, eating breakfast; medium-shot; Normal People; A cozy kitchen with warm, natural light streaming through the window; cinematic
Characteristic
Shot : A young woman is eating breakfast in a kitchen, with warm sunlight streaming in through the window. She is sitting at a table, her face half-turned away from the camera. The light is soft and diffused, casting long shadows across the table and the woman’s face. The scene is simple but evocative, with a sense of quiet contemplation.
Aesthetic Score : 0.7
Mood : calm, contemplative, cozy
Quality
Entropy : 6.33
Noise : 86
Prompt Clip Score : 0.32
AI Evaluation
Likelihood of AI : 0.20
Image errors : No visible errors
Lost in the Shadows
A solitary figure walks through the night, their shadow stretching long behind them. The single street lamp casts a melancholic glow, creating a sense of mystery and isolation.
Prompt
key-lighting Low-key lighting with a single, harsh light source: Lonely, suspenseful, eerie ; A man walking down a dark, deserted street; medium-shot; Single Person; A streetlamp casting a pool of light on the pavement; cinematic
Characteristic
Shot : A single figure walks in the beam of a streetlight in a nighttime scene.
Aesthetic Score : 0.7
Mood : lonely, somber, mysterious
Quality
Entropy : 4.95
Noise : 93
Prompt Clip Score : 0.25
AI Evaluation
Likelihood of AI : 0.10
Image errors : No visible artifacts or errors.
Shadows and Secrets: A Gathering by Firelight
Five figures huddle around a crackling campfire, their faces illuminated by the dancing flames. The smoky atmosphere and long shadows cast on the cave walls create a sense of mystery and intrigue, hinting at a story waiting to be told. This image evokes a cozy yet contemplative mood, leaving you wondering what secrets these men hold.
Prompt
key-lighting Firelight with warm, flickering shadows: Warm, friendly, nostalgic ; A group of friends gathered around a campfire; studio; Normal People; A smoky, firelit scene with the flames reflecting in their faces; cinematic
Characteristic
Shot : A group of five men are sitting around a campfire in a dark, smoky cave.
Aesthetic Score : 0.7
Mood : brooding, mysterious, masculine
Quality
Entropy : 6.06
Noise : 114
Prompt Clip Score : 0.29
AI Evaluation
Likelihood of AI : 0.10
Image errors : There are no visible artifacts or errors in the image.
Lost in the Shadows
A solitary figure stands in a vast, empty room, bathed in the ethereal glow of a single window. The play of light and shadow creates a sense of mystery and isolation, leaving the viewer to ponder the man’s secrets and the weight of his solitude.
Prompt
key-lighting Low-key lighting with a single, harsh light source: Intense, suspenseful, mysterious ; A detective standing in a dimly lit interrogation room; medium-shot; Hero; A room with a single overhead light casting harsh shadows; cinematic
Characteristic
Shot : A man in a suit stands in a dark, industrial space illuminated by a single window. Sunbeams stream through the window, casting dramatic shadows.
Aesthetic Score : 0.7
Mood : mysterious, brooding, cinematic
Quality
Entropy : 6.22
Noise : 109
Prompt Clip Score : 0.26
AI Evaluation
Likelihood of AI : 0.10
Image errors : The image is slightly grainy and the shadows appear somewhat artificial.
Golden Hour Serenity
A young woman finds peace and tranquility amidst the warm, golden light filtering through the trees in a serene park setting. Her contemplative gaze adds to the overall sense of calm.
Prompt
key-lighting Natural lighting with soft, diffused light: Melancholy, contemplative, peaceful ; A woman sitting on a park bench, lost in thought; medium-shot; Single Person; A park with dappled sunlight filtering through the trees; cinematic
Characteristic
Shot : A young woman sits on a park bench, looking off to the side. The sun is shining brightly through the trees, creating a warm, inviting atmosphere.
Aesthetic Score : 0.7
Mood : serene, contemplative, peaceful
Quality
Entropy : 6.61
Noise : 102
Prompt Clip Score : 0.23
AI Evaluation
Likelihood of AI : 0.10
Image errors : There are no major image errors, but some mild artifacts from compression are visible on the leaves.
Silhouetted Against the Setting Sun: A Soldier’s Contemplation
A lone soldier stands in silhouette against a breathtaking sunset, the hazy field before him adding to the sense of isolation and contemplation. The image evokes a melancholic and powerful mood, capturing the weight of the soldier’s experience.
Prompt
key-lighting Backlighting with a strong, warm light source: Epic, heroic, hopeful ; A soldier standing on a battlefield, silhouetted against the rising sun; medium-shot; Hero; A desolate landscape with a fiery sunrise; cinematic
Characteristic
Shot : A silhouette of a soldier standing in a field at sunset.
Aesthetic Score : 0.6
Mood : solitude, melancholic, dramatic
Quality
Entropy : 6.27
Noise : 98
Prompt Clip Score : 0.29
AI Evaluation
Likelihood of AI : 0.80
Image errors : The image has some blurriness, particularly in the background.
Intimate Family Dinner Under a Warm Glow
A family of four gathers around a dimly lit kitchen table, bathed in the soft light of a pendant lamp. The cozy atmosphere is enhanced by the dark silhouettes of trees outside the windows, adding a touch of mystery to the scene.
Prompt
key-lighting Soft, diffused lighting with warm tones: Happy, cozy, intimate ; A family gathered around a dinner table; studio; Normal People; A warm, inviting kitchen with soft, overhead lighting; cinematic
Characteristic
Shot : A family is gathered around a table in a kitchen, eating dinner. The scene is dimly lit, with only a single overhead light illuminating the table. The kitchen is fairly empty and minimally decorated.
Aesthetic Score : 0.6
Mood : cozy, warm, intimate
Quality
Entropy : 5.74
Noise : 95
Prompt Clip Score : 0.28
AI Evaluation
Likelihood of AI : 0.60
Image errors : The image has a slight graininess, which might be due to the low lighting or digital noise. The shadows are somewhat harsh and artificial, which makes the image look a bit too staged.
Unveiling the Secrets: A Scientist’s Focused Pursuit in a Glowing Lab
A scientist, bathed in the eerie glow of chemical reactions, works intently in a dimly lit lab. Their focused expression and the mysterious luminescence create a sense of intrigue and anticipation, hinting at groundbreaking discoveries in the making.
Prompt
key-lighting High-key lighting with a mix of bright, focused lights and dark shadows: Intriguing, mysterious, futuristic ; A scientist working in a laboratory, surrounded by glowing equipment; medium-shot; Hero; A laboratory with a mix of bright, focused lights and dark shadows; cinematic
Characteristic
Shot : A scientist in a lab setting, working with equipment in a dark, illuminated space.
Aesthetic Score : 0.7
Mood : serious, focused, industrial
Quality
Entropy : 5.47
Noise : 98
Prompt Clip Score : 0.18
AI Evaluation
Likelihood of AI : 0.20
Image errors : The image has some minor noise and grain, which is typical of low-light photography.
Conclusion
The results show that the generative AI model performed well in terms of understanding the scene and aesthetics, but struggled with camera positioning. Here’s a breakdown:
- Camera Position: The model scored 0.1, indicating a significant difference between the intended camera position in the prompt and the actual camera position in the generated image. This suggests the model is not very good at following camera position instructions.
- Shot Analysis: The model scored 0.52, which is considered good. This means the model was able to understand the scene in the prompt and create an image that reflects it reasonably well.
- Aesthetic Analysis: The model scored 0.1, which is considered very good. This indicates that the generated image closely matches the expected aesthetic style.
Overall: The model demonstrates a strong ability to understand the scene and create aesthetically pleasing images, but it needs improvement in accurately interpreting camera position instructions.