Research map / Learning theory and algorithms / Matrix and tensor methods
A Novel Geometric Framework on Gram Matrix Trajectories for Human Behavior Understanding
Anis Kacem, Mohamed Daoudi, Boulbaba Ben Amor, Stefano Berretti, Juan Carlos Alvarez-Paiva
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add spatial covariance (Gram matrix on S+) to Grassmann shape trajectories (ablation) | landmark trajectory classifier · UT-Kinect | accuracy (%): 92.46 -> 96.48 | Improved |
| add spatial covariance (Gram matrix on S+) to Grassmann shape trajectories (ablation) | landmark trajectory classifier · Florence3D | accuracy (%): 75 ± 5.22 -> 88.07 ± 4.8 | Improved |
| add spatial covariance (Gram matrix on S+) to Grassmann shape trajectories (ablation) | landmark trajectory classifier · SBU Interaction | accuracy (%): 76.3 ± 3.26 -> 88.45 ± 2.88 | Improved |
| add spatial covariance (Gram matrix on S+) to Grassmann shape trajectories (ablation) | landmark trajectory classifier · P-BME emotion recognition | accuracy (%): 66.35 ± 6.43 -> 78.15 ± 5.79 | Improved |
| add dynamic time warping alignment (ablation) | landmark trajectory classifier · UT-Kinect | accuracy (%): 91.46 -> 96.48 | Improved |
| add dynamic time warping alignment (ablation) | landmark trajectory classifier · SBU Interaction | accuracy (%): 81.36 ± 2.78 -> 88.45 ± 2.88 | Improved |
| add dynamic time warping alignment (ablation) | landmark trajectory classifier · P-BME emotion recognition | accuracy (%): 63.23 ± 8.62 -> 78.15 ± 5.79 | Improved |
| use closeness on S+ instead of flat Frobenius distance (ablation) | trajectory comparison · UT-Kinect | accuracy (%): 92.96 -> 96.48 | Improved |
| use closeness on S+ instead of regularized distance on positive definite cone (ablation) | trajectory comparison · UT-Kinect | accuracy (%): 94.98 -> 96.48 | Improved |
| use ppfSVM instead of k-NN (ablation) | S+ trajectory classifier · SBU Interaction | accuracy (%): 61.06 ± 2.3 -> 88.45 ± 2.88 | Improved |
| add late fusion of body-part classifiers (ablation) | S+ trajectory classifier · SBU Interaction | accuracy (%): 88.45 ± 2.88 -> 93.7 ± 1.59 | Improved |
| add late fusion of body-part classifiers (ablation) | S+ trajectory classifier · SYSU-3D | accuracy (%): 76.01 ± 2.09 -> 80.22 ± 2.09 | Improved |
Closest papers by meaning
- A Novel Space-Time Representation on the Positive Semidefinite Cone for Facial Expression Recognition (ICCV 2017 · similarity 0.93)
- Sparse Coding of Shape Trajectories for Facial Expression and Action Recognition (TPAMI 2019 · similarity 0.90)
- Elastic Functional Coding of Riemannian Trajectories (TPAMI 2016 · similarity 0.88)
- Geometry-Aware Similarity Metrics for Neural Representations on Riemannian and Statistical Manifolds (NeurIPS 2026 · similarity 0.86)
- A general kernel framework for non-CND distance measures using $|\mathcal{D}|$-dimensional sparse landmark embeddings (NeurIPS 2026 · similarity 0.84)
- Task-Induced Riemannian Metrics for Vision Transformer Feature Spaces (NeurIPS 2026 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
