Dark Mode
Fake and Real Brand Logo Dataset
Data Science and Analytics
Tags and Keywords
Trusted By
"No reviews yet"
Free
About
This dataset is designed for the detection of fake logos using semantic similarity. It comprises logos that have been resized to a uniform shape of 70x70 pixels, which helps to reduce computational demands and model complexity during training. The original files and the code utilised for data mining and processing are available on a GitHub repository.
Columns
The dataset primarily includes logo image data and associated classifications. While specific column headers from a sample file are not detailed in the provided information, a typical structure for this dataset would involve:
- Logo Image Data: The visual representation of the logo, likely in a 70x70 pixel format.
- Logo Label: A classification indicating whether the logo is 'Fake' or 'Real'.
- Brand: The name of the brand associated with the logo (e.g., Samsung, Pepsi, Nike), derived from a list of legitimate brands.
Distribution
The dataset is available in a version 5 release, with a total size of 2.26 MB. It includes various file types, such as a
Logos.txt file (481 B) which lists numerous brand names. Other supporting files mentioned are genLogoOutput, output, and file_mapping.csv. The core data consists of images structured for efficient processing, and specific numbers for rows or records of image files are not explicitly stated.Usage
This dataset is ideal for computer vision applications, particularly in the domain of object detection and image classification. It can be used to:
- Develop and train machine learning models for identifying fake logos.
- Implement solutions for brand protection and authenticity verification.
- Conduct research in visual fraud detection.
- Facilitate training with popular machine learning frameworks such as TensorFlow and Keras.
Coverage
The dataset's coverage includes logos from a wide array of well-recognised brands such as Bic, Samsung, Pepsi, Lays, Nike, Google, Apple, and many others. It does not specify particular geographic regions, time ranges, or demographic scopes, focusing solely on the visual characteristics of the logos themselves.
License
CC0: Public Domain
Who Can Use It
This dataset is primarily intended for:
- Machine learning engineers and data scientists working on image classification and object detection tasks.
- Researchers focusing on computer vision, brand authenticity, and digital forensics.
- Developers building applications that require logo recognition and verification capabilities.
Dataset Name Suggestions
- Logo Authenticity Verification Dataset
- Fake and Real Brand Logo Dataset
- Digital Logo Detection Dataset
- Resized Logo Classification Set
- Brand Logo Integrity Dataset
Attributes
Original Data Source: Fake and Real Brand Logo Dataset
QUALITY