BVC-Face Dataset
A dataset of face images collected by the Biometrics Visions Computing research group for biometric and computer vision research.
- Face image dataset
- Collected by BVC
- Biometric recognition research
RESEARCH RESOURCES
Datasets & Software from BVC
Biometrics Visions Computing has developed research datasets, algorithms and computational resources supporting work in biometrics, computer vision, image processing and artificial intelligence.
These resources have been created through BVC research activities and are intended to support reproducible scientific investigation, algorithm development and comparative research.
BVC DATASETS
Research Datasets
BVC-Body Dataset
A dataset of body images collected by BVC for research involving body-based biometrics, soft biometrics and person recognition.
- Body image dataset
- Collected by BVC
- Soft biometric research
BVC Gender & Age from Voice Challenging Dataset
A challenging voice dataset developed for research involving voice biometrics, gender and age analysis, and speaker-related classification.
- Voice utterances
- Challenging recording conditions
- Gender and age research
BVC Afro Voice Dataset
A multilingual voice biometric dataset containing English and vernacular utterances collected from African subjects.
- 94 subjects
- 940 voice prints
- English and vernacular speech
FEATURED DATASET
BVC Afro Voice Dataset
The BVC Afro Voice Dataset contains clear voice prints acquired from 94 subjects. Ten voice utterances were recorded for each subject: five spoken in English and five spoken in vernacular, producing a total of 940 voice prints.
The dataset supports research in multilingual voice biometrics and recognition in contexts where speakers may use both English and indigenous languages.
94Subjects
10Utterances per subject
940Voice prints
ALGORITHMS & SOFTWARE
Fingerprint Preprocessing & Binarization
BVC research has produced an adaptive fingerprint preprocessing and binarization algorithm designed for fair- and poor-quality plain and rolled fingerprints, including images containing large regions of degraded quality.
The approach performs adaptive contrast enhancement and local thresholding before orientation-field estimation, helping to enhance smudged and faded ridge structures in recoverable fingerprint regions.
OPEN SOURCE
Fingerprint Process, Enhance & Binarize
Source code for fingerprint enhancement and binarization is available through GitHub.
View on GitHub →DATASET ACCESS
Requesting BVC Datasets
Some BVC datasets are available for research use subject to appropriate access conditions and acknowledgement requirements. Researchers interested in using a dataset may contact BVC with information about their proposed research and the resource they wish to access.
Contact BVC