Scene Completion Using Millions of Photographs
E326788
"Scene Completion Using Millions of Photographs" is a seminal computer vision and graphics paper that introduced a data-driven method for automatically filling in missing regions of images by searching a massive online photo collection for visually compatible patches.
All labels observed (2)
| Label | Occurrences |
|---|---|
| Scene Completion Using Millions of Photographs canonical | 2 |
| content-aware fill in Adobe Photoshop | 1 |
How this entity was disambiguated
This entity first appeared as the object of triple T3094204 — resolving that mention is where its identity was fixed. The disambiguator weighed these candidate entities and picked the highlighted one (or “None”, minting a new entity). This is how homonymy is resolved: the same surface form can point to different entities.
NED1
Entity disambiguation (via context triple)
gpt-5-mini-2025-08-07
Target entity: Scene Completion Using Millions of Photographs Context triple: [Alexei Efros, notableWork, Scene Completion Using Millions of Photographs]
-
A.
Long-term Recurrent Convolutional Networks for Visual Recognition and Description
"Long-term Recurrent Convolutional Networks for Visual Recognition and Description" is a research paper that introduces a deep learning architecture combining convolutional and recurrent neural networks to perform tasks like video recognition and automatic image or video captioning.
-
B.
Modeling image patches with a directed hierarchy of Markov random fields
"Modeling image patches with a directed hierarchy of Markov random fields" is a research paper that introduces a probabilistic hierarchical model for capturing complex statistical structure in image patches using directed Markov random fields.
-
C.
Show and Tell: A Neural Image Caption Generator
"Show and Tell: A Neural Image Caption Generator" is a pioneering deep learning model that automatically generates natural-language descriptions for images by combining convolutional and recurrent neural networks.
-
D.
Look Around street-level imagery
Look Around street-level imagery is Apple Maps’ interactive, high-resolution street-view experience that lets users virtually explore streets and surroundings in a seamless, panoramic way.
-
E.
Markov random fields
Markov random fields are probabilistic graphical models that represent the joint distribution of a set of random variables with local dependencies encoded by an undirected graph, widely used in areas like statistical physics, computer vision, and spatial statistics.
- F. None of above. chosen
- G. Unsure - the case is ambiguous/there is not enough information to decide.
NED2
Entity disambiguation (via description)
gpt-5-mini-2025-08-07
Target entity: Scene Completion Using Millions of Photographs Target entity description: "Scene Completion Using Millions of Photographs" is a seminal computer vision and graphics paper that introduced a data-driven method for automatically filling in missing regions of images by searching a massive online photo collection for visually compatible patches.
-
A.
Long-term Recurrent Convolutional Networks for Visual Recognition and Description
"Long-term Recurrent Convolutional Networks for Visual Recognition and Description" is a research paper that introduces a deep learning architecture combining convolutional and recurrent neural networks to perform tasks like video recognition and automatic image or video captioning.
-
B.
Modeling image patches with a directed hierarchy of Markov random fields
"Modeling image patches with a directed hierarchy of Markov random fields" is a research paper that introduces a probabilistic hierarchical model for capturing complex statistical structure in image patches using directed Markov random fields.
-
C.
Show and Tell: A Neural Image Caption Generator
"Show and Tell: A Neural Image Caption Generator" is a pioneering deep learning model that automatically generates natural-language descriptions for images by combining convolutional and recurrent neural networks.
-
D.
Look Around street-level imagery
Look Around street-level imagery is Apple Maps’ interactive, high-resolution street-view experience that lets users virtually explore streets and surroundings in a seamless, panoramic way.
-
E.
Markov random fields
Markov random fields are probabilistic graphical models that represent the joint distribution of a set of random variables with local dependencies encoded by an undirected graph, widely used in areas like statistical physics, computer vision, and spatial statistics.
- F. None of above. chosen
Statements (45)
| Predicate | Object |
|---|---|
| instanceOf |
computer graphics paper
ⓘ
computer vision paper ⓘ research paper ⓘ |
| application |
filling in occluded regions
ⓘ
photo editing ⓘ removing unwanted objects from photos ⓘ |
| approach |
data-driven method
ⓘ
example-based method ⓘ |
| assumption |
similar scenes exist in large online photo collections
ⓘ
visually compatible patches can be found in a massive dataset ⓘ |
| category |
content-aware image editing methods
ⓘ
image-based rendering techniques ⓘ scene understanding methods ⓘ |
| coreIdea |
fill missing regions of images using patches from a large photo collection
ⓘ
search a massive online photo collection for visually compatible patches ⓘ use real-world photographs to complete scenes in a target image ⓘ |
| datasetSource |
Internet image repositories
ⓘ
online photo collections ⓘ |
| field |
computer graphics
ⓘ
computer vision ⓘ image processing ⓘ |
| goal |
automatically fill in missing regions of images
ⓘ
produce visually plausible scene completions ⓘ |
| impact |
influential in example-based image editing
ⓘ
inspired later work on large-scale visual data for graphics ⓘ seminal work in data-driven image completion ⓘ widely cited in computer vision and graphics literature ⓘ |
| matchingCriterion |
context similarity
ⓘ
visual compatibility ⓘ |
| novelty |
demonstrated data-driven scene completion using millions of photographs
ⓘ
introduced large-scale Internet photo collections into image completion ⓘ |
| researchArea |
data-driven graphics
ⓘ
example-based image synthesis ⓘ image completion ⓘ image inpainting ⓘ patch-based image editing ⓘ |
| task |
hole filling in images
ⓘ
object removal from photographs ⓘ scene completion ⓘ |
| title | Scene Completion Using Millions of Photographs ⓘ |
| uses |
context-aware patch selection
ⓘ
large-scale image database ⓘ millions of Internet photographs ⓘ patch-based matching ⓘ visual similarity search ⓘ |
How these facts were elicited
The pipeline generated the facts above by prompting gpt-5.1 with this entity's name + description and the instruction below.
Instruction
You are a knowledge base construction expert. Given a subject entity and a description of it, return factual statements that you know for the subject as a JSON list of dictionaries(triples), where keys must be "subject", "predicate" and "object". The number of facts may be very high, between 25 to 50 or more, for very popular subjects. For less popular subjects, the number of facts can be very low, like 5 or 10. # Requirements - If you don't know the subject at all, return an empty list. - If the subject is not a named entity, return an empty list. - Include at least one triple where predicate is "instanceOf". - Do not get too wordy. - Separate several objects into multiple triples with one object.
Input
Subject: Scene Completion Using Millions of Photographs Description of subject: "Scene Completion Using Millions of Photographs" is a seminal computer vision and graphics paper that introduced a data-driven method for automatically filling in missing regions of images by searching a massive online photo collection for visually compatible patches.
Referenced by (3)
Full triples — surface form annotated when it differs from this entity's canonical label.
Scene Completion Using Millions of Photographs
→
title
→
Scene Completion Using Millions of Photographs
ⓘ
subject linked to:
Adobe Sensei AI platform
linked to: Scene Completion Using Millions of Photographs