Grouped-People Splitting Based on Face Detection and Body Proportion Constraints

Piva, S.; Comes, L.; Asadi, M.; Regazzoni, Carlo S.

doi:10.1109/avss.2006.56

Cited by 1 publication

(2 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The bounding box was enlarged to include more features resulting from the face detector. The length of the idealized human figure is approximately eight heads tall divided as follows [29] From the second point, we can deduce that the bottom of the bounding box has to be extended with the same length as the face/head to include the neck, shoulders, and part of the human torso. To cover part of the hair, we extended the bounding box from the top to the highest coordinate in the image.…”

Section: Extending Featuresmentioning

confidence: 99%

See 1 more Smart Citation

Multi-Modal Long-Term Person Re-Identification Using Physical Soft Bio-Metrics and Body Figure

2022

View full text Add to dashboard Cite

Person re-identification is the task of recognizing a subject across different non-overlapping cameras across different views and times. Most state-of-the-art datasets and proposed solutions tend to address the problem of short-term re-identification. Those models can re-identify a person as long as they are wearing the same clothes. The work presented in this paper addresses the task of long-term re-identification. Therefore, the proposed model is trained on a dataset that incorporates clothes variation. This paper proposes a multi-modal person re-identification model. The first modality includes soft bio-metrics: hair, face, neck, shoulders, and part of the chest. The second modality is the remaining body figure that mainly focuses on clothes. The proposed model is composed of two separate neural networks, one for each modality. For the first modality, a two-stream Siamese network with pre-trained FaceNet as a feature extractor for the first modality is utilized. Part-based Convolutional Baseline classifier with a feature extractor network OSNet for the second modality. Experiments confirm that the proposed model can outperform several state-of-the-art models achieving 81.4 % accuracy on Rank-1, 82.3% accuracy on Rank-5, 83.1% accuracy on Rank-10, and 83.7% accuracy on Rank-20.

show abstract

Section: Extending Featuresmentioning

confidence: 99%

“…The extension will not only include regular hairstyles but also extreme hairstyles (e.g., high ponytails), hijabs, and turbans. According to [29], the width of the idealized human head is 2/3 times its height. Consequently, the width of the bounding box was extended with 70% of the length of the head to include the shoulders.…”

Section: Extending Featuresmentioning

confidence: 99%