Gait Energy Image Features for Human Biometrics

Version 24.3 of a dataset containing 195 Gait Energy Image (GEI) features for human biometrics. The data is derived from CNN processing and is categorized under computer vision. It is hosted on Kaggle, but specific authorship and temporal details are not provided.

Use Cases

• Train a person re-identification model using the 195-dimensional GEI feature vectors.
• Develop a gait analysis classifier for activity recognition based on the CNN-extracted features.
• Benchmark biometric verification algorithms using the provided GEI feature embeddings.

Strengths

• Contains 195 distinct features per sample, providing a detailed representation.
• Features are derived from Gait Energy Images, a standard representation in biometrics.

Limitations

• The total number of samples (rows) is unknown, making it difficult to assess scale.
• The dataset's origin, collection methodology, and potential demographic or environmental biases are unspecified.

Provenance

Source

Kaggle

Collection Method

Features likely extracted from video or image sequences using a Convolutional Neural Network (CNN).