RECORD
Seeing Faces in Things: A Model and Dataset for Pareidolia
Computer Vision and Pattern Recognition · Submitted 24 Sep 2024
abstract compiled · 2026-09-23 · no language-model review
The authors' main claim Through studies on human subjects and our pareidolic face detectors we confirm a key prediction of our model regarding what image conditions are most likely to induce pareidolia.
The compiled record
The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.
Open the compiled record
The abstract, sentence by sentence
The human visual system is well-tuned to detect faces of all shapes and sizes. While this brings obvious survival advantages, such as a better chance of spotting unknown predators in the bush, it also leads to spurious face detections. ``Face pareidolia'' describes the perception of face-like structure among otherwise random stimuli: seeing faces in coffee stains or clouds in the sky. In this paper, we study face pareidolia from a computer vision perspective. We present an image dataset of ``Faces in Things'', consisting of five thousand web images with human-annotated pareidolic faces. Using this dataset, we examine the extent to which a state-of-the-art human face detector exhibits pareidolia, and find a significant behavioral gap between humans and machines. We find that the evolutionary need for humans to detect animal faces, as well as human faces, may explain some of this gap. Finally, we propose a simple statistical model of pareidolia in images. Through studies on human subjects and our pareidolic face detectors we confirm a key prediction of our model regarding what image conditions are most likely to induce pareidolia. Dataset and Website: https://aka.ms/faces-in-things
background contribution method result interpretation main claim
| # | sentence | role | the authors' own work | states the main conclusion | strength |
|---|---|---|---|---|---|
| s0 | The human visual system is well-tuned to detect faces of all shapes and sizes. | background 1.00 | 0.05 | 0.02 | asserted 1.00 |
| s1 | While this brings obvious survival advantages, such as a better chance of spotting unknown predators in the bush, it also leads to spurious face detections. ``Face pareidolia'' describes the perception of face-like structure among otherwise random stimuli: seeing faces in coffee stains or clouds in the sky. | background 1.00 | 0.04 | 0.04 | asserted 0.94 |
| s2 | In this paper, we study face pareidolia from a computer vision perspective. | contribution 0.81 | 0.70 | 0.02 | asserted 0.84 |
| s3 | We present an image dataset of ``Faces in Things'', consisting of five thousand web images with human-annotated pareidolic faces. | contribution 0.73 | 0.97 | 0.05 | asserted 1.00 |
| s4 | Using this dataset, we examine the extent to which a state-of-the-art human face detector exhibits pareidolia, and find a significant behavioral gap between humans and machines. | result 0.94 | 0.97 | 0.36 | asserted 1.00 |
| s5 | We find that the evolutionary need for humans to detect animal faces, as well as human faces, may explain some of this gap. | interpretation 0.99 | 0.66 | 0.17 | hedged possibility 1.00 |
| s6 | Finally, we propose a simple statistical model of pareidolia in images. | contribution 0.72 | 0.94 | 0.08 | asserted 0.98 |
| s7 | Through studies on human subjects and our pareidolic face detectors we confirm a key prediction of our model regarding what image conditions are most likely to induce pareidolia. | result 0.89 | 0.95 | 0.27 | asserted 1.00 |
| s8 | Dataset and Website: https://aka.ms/faces-in-things | contribution 0.96 | 0.40 | 0.02 | unclear 0.76 |
What was examined, and what was not
- arXiv listing abstract v1
- full text
- figures and tables
- references and citing sentences
- relations to other papers
- P abstract
- S sections
- T body sentences
- C citations
- R cross-paper relations
The paper
Original abstract
The human visual system is well-tuned to detect faces of all shapes and sizes. While this brings obvious survival advantages, such as a better chance of spotting unknown predators in the bush, it also leads to spurious face detections. ``Face pareidolia'' describes the perception of face-like structure among otherwise random stimuli: seeing faces in coffee stains or clouds in the sky. In this paper, we study face pareidolia from a computer vision perspective. We present an image dataset of ``Faces in Things'', consisting of five thousand web images with human-annotated pareidolic faces. Using this dataset, we examine the extent to which a state-of-the-art human face detector exhibits pareidolia, and find a significant behavioral gap between humans and machines. We find that the evolutionary need for humans to detect animal faces, as well as human faces, may explain some of this gap. Finally, we propose a simple statistical model of pareidolia in images. Through studies on human subjects and our pareidolic face detectors we confirm a key prediction of our model regarding what image conditions are most likely to induce pareidolia. Dataset and Website: https://aka.ms/faces-in-things
Human review
How the literature uses it
What this paper does with the arXiv papers it cites
What it cites · 3 sentences without a confident label
Uses the method or tool of 2 papers
-
We perform experiments using both their MobileNet [ref] and ResNet50 [ref] backbones and use the Pytorch_Retinaface [ref] repository to ensure the same experimental conditions, dataset characteristics, and preprocessing.
MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications · Experiments · 1.00
-
We use pre-trained models provided by this repository and fine-tune them for 10 epochs with the AdamW optimizer [ref] using a learning rate of $10^-4$ and a weight decay of $5×10^-4$.
Decoupled Weight Decay Regularization · Experiments · 0.99
Cites as background 1 paper
-
This has been explored particularly in the context of detecting faces, people and objects across modalities such as photography, different art movements, cartoons and sketches [ref] .
The Cross-Depiction Problem: Computer Vision Algorithms for Recognising Objects in Artwork and in Photographs · Related Work · 0.93
Discussion
Know something about this paper? Leave a comment, or highlight any passage above to comment on it.