The Economist Historical Advertisements - Faces Dataset

Creators:

Kluge, Stefan

Publication Date:

2023

Data Category:

Dataset Description:

This dataset contains 116.746 identified faces (bounding box location on image, predicted age and gender) for all historical advertisements from all 8,840 issues of The Economist magazine, years 1843 to 2014. Faces have been detected using the following library: https://pythonrepo.com/repo/timesler-facenet-pytorch-python-deep-learning. You will need the The Economist Historical Advertisements - Master Dataset as well, to work with the data. In total, the dataset has a size of 20,2 MB and is organized as follows:

Filename: A unique identifier corresponding to each advertisement where a face has been detected. This identifier links directly to the specific advertisement within The Economist archives.
Bounding Box Coordinates:
- Bounding Box relative X1 and Y1: These values represent the top-left corner coordinates of the bounding box encapsulating the detected face, expressed as proportions relative to the image dimensions.
- Bounding Box relative X2 and Y2: These values denote the bottom-right corner coordinates of the bounding box, also as relative proportions. To determine the absolute pixel coordinates, multiply these relative values by the image's width and height, respectively.
Segmentation Confidence Score: A numerical value indicating the confidence level of the neural network algorithm that the identified bounding box indeed contains a face. Higher scores reflect greater confidence in accurate face detection.
Size Relative: A metric indicating the proportion of the advertisement occupied by the detected face. For example, a value of 1 signifies that the face covers the entire advertisement, while 0.5 indicates it covers half.
Predicted Age: An estimated age of the individual based on facial analysis performed by the detection algorithm.
Gender Probability: A probability score representing the likelihood of the detected face being female. A value of 0 indicates male, 1 indicates female, and intermediate values (e.g., 0.4) suggest a 40% likelihood of the individual being female

Variables:

Name	Description
Filename	Unique identifier of the advertisement this face appears on
Bounding Box relative X1	Left-top coordinate of a rectangle identifying the face on the page, relative to the pixel coordinates of the image from column 2 ("URLs …") of the Master Dataset (which is related to this dataset by the unique identifier in column 1). Multiply this value by the width of the image to get the absolute x coordinate. If the ad is a multi page ad, the images from column 2 have to be horizontally concatenated first.
Bounding Box relative Y1	Left-top coordinate
Bounding Box relative X2	Right-bottom coordinate
Bounding Box relative Y2	Right-bottom coordinate
Segmentation confidence score	Confidence of the neural network algorithm that these bounding boxes represent a face.
Size relative	1 = Face covers all of the ad; 0.5 = Face covers half the ad.
Age	Predicted age of the face.
Gender	Gender probability; 0 = male; 1 = female; 0.4 = 40% likelyhood of beeing female

Details:

Bookmark this Dataset/Publication

The Economist Historical Advertisements - Faces Dataset

ConceptNet

IMDb movies extensive dataset

Debt securities statistics

The Economist Historical Advertisements - Faces Dataset

Sign In

Register

Reset Password