Photos | Sushi Night Out with Friends
Fuminori Komatsu and Bruce K joined seven others for a delicious meal at a sushi restaurant in 2002.
BLIP-2 Description:
a group of people posing for a photo at a sushi restaurantMetadata
Capture date:
Original Dimensions:
640w x 480h - (download 4k)
Usage
Dominant Color:
restaurant accessories architecture urban food table shoe cap lighting vaga bruce wristwatch electronics bride cafeteria wedding respect james couch k baby alcohol interior junglescene furniture indoors hat dining sterling chair aurakul mike momo room bar building photography zona pictures night cup portrait trace pants club plate fuminori komatsu party footwear meal art bbq painting grilling life speaker cooking pub dish baseball jewelry bracelet beverage counter fun glasses
Detected Text
overall
(31.15%)
curation
(65.50%)
highlight visibility
(5.54%)
behavioral
(70.64%)
failure
(-0.27%)
harmonious color
(-2.49%)
immersiveness
(0.07%)
interaction
(1.00%)
interesting subject
(5.47%)
intrusive object presence
(-12.28%)
lively color
(-2.17%)
low light
(69.38%)
noise
(-17.43%)
pleasant camera tilt
(-5.40%)
pleasant composition
(-61.62%)
pleasant lighting
(-33.91%)
pleasant pattern
(4.20%)
pleasant perspective
(5.63%)
pleasant post processing
(2.29%)
pleasant reflection
(1.54%)
pleasant symmetry
(0.24%)
sharply focused subject
(0.68%)
tastefully blurred
(-3.79%)
well chosen subject
(-42.55%)
well framed subject
(1.71%)
well timed shot
(11.81%)
all
(-4.46%)
* NOTE: Amazon Rekognition
detected a celebrity in this image using the
Celebrity Recognition API. The API isn't perfect, but it does give you the MatchConfidence which I display
next to the celebrity's name along with links _↗ to their info.
* NOTE: This image was scaled up from its original size using an AI model called GFP-GAN (Generative Facial Prior), which is a
Generative adversartial network that can be used to repair (or upscale in this case) photos, sometimes the results are a little...
weird.
* WARNING: The title and caption of this image were generated with AI (gpt-3.5-turbo-0301
from
OpenAI) based on a
BLIP-2 image-to-text labeling, tags,
location,
people
and album metadata from the image and are
potentially inaccurate, often hilariously so. If you'd like me to adjust anything,
just reach out.