Posenet-Mobilenet: Pose Estimation
PoseNet-MobileNet is a lightweight human pose estimation model that combines the PoseNet algorithm with the MobileNet backbone, designed for efficient real-time keypoint detection on mobile and edge devices. The model can predict the positions of human body keypoints such as the head, shoulders, elbows, and knees, and is suitable for both single-person and multi-person pose estimation tasks. Leveraging MobileNet’s efficient feature extraction, PoseNet-MobileNet achieves a good balance between accuracy and speed with low computational cost. It is widely used in applications such as fitness tracking, augmented reality, and human-computer interaction.
Source model
- Input shape: 513x513
- Number of paramaters: 7.63M
- Model size: 13.0M
- Output shape: 1x17x33x33,1x34x33x33,1x32x33x33,1x32x33x33
Source model repository: Posenet-Mobilenet
Performance Reference
Please search model by model name in Model Farm
Inference & Model Conversion
Please search model by model name in Model Farm
License
Source Model: APACHE-2.0
Deployable Model: APLUX-MODEL-FARM-LICENSE