Research map / Video / Video segmentation and tracking
A Two Stream Siamese Convolutional Neural Network for Person Re-Identification
Dahjung Chung, Khalid Tahboub, Edward J. Delp
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 the temporal stream to the spatial stream (ablation) | two-stream Siamese CNN · PRID2011 | CMC rank-1 accuracy (%): 75 -> 78 | Improved |
| add the temporal stream to the spatial stream (ablation) | two-stream Siamese CNN · iLIDS-VID | CMC rank-1 accuracy (%): 57 -> 58 | Improved |
| raise spatial weight from 1 to 2 with temporal weight 1 (ablation) | two-stream Siamese CNN · iLIDS-VID | CMC rank-1 accuracy (%): 58 -> 60 | Improved |
| raise spatial weight from 2 to 3 with temporal weight 1 (ablation) | two-stream Siamese CNN · iLIDS-VID | CMC rank-1 accuracy (%): 60 -> 56 | Hurt |
| raise spatial weight from 2 to 3 with temporal weight 1 (ablation) | two-stream Siamese CNN · PRID2011 | CMC rank-1 accuracy (%): 78 -> 79 | Improved |
| lengthen probe/gallery test sequences from 16 to 32 frames | two-stream Siamese CNN · iLIDS-VID | CMC rank-1 accuracy (%): 41 -> 50 | Improved |
| lengthen probe/gallery test sequences from 16 to 128 frames | two-stream Siamese CNN · iLIDS-VID | CMC rank-1 accuracy (%): 41 -> 58 | Improved |
| lengthen test sequences from 64 to 128 frames | two-stream Siamese CNN · iLIDS-VID | CMC rank-1 accuracy (%): 56 -> 58 | Improved |
| use proposed method instead of GOG+XQDA | two-stream Siamese CNN · PRID2011 | CMC rank-1 accuracy (%): 74 -> 78 | Improved |
| use proposed method instead of RNN-ReID | two-stream Siamese CNN · iLIDS-VID | CMC rank-1 accuracy (%): 58 -> 60 | No clear effect |
Closest papers by meaning
- Gated Siamese Convolutional Neural Network Architecture for Human Re-identification (ECCV 2016 · similarity 0.90)
- Recurrent Convolutional Network for Video-Based Person Re-Identification (CVPR 2016 · similarity 0.89)
- Person Search via A Mask-guided Two-stream CNN Model (ECCV 2018 · similarity 0.89)
- Re-Identification With Consistent Attentive Siamese Networks (CVPR 2019 · similarity 0.89)
- Deep Neural Networks with Inexact Matching for Person Re-Identification (NeurIPS 2016 · similarity 0.89)
- Efficient and Deep Person Re-Identification Using Multi-Level Similarity (CVPR 2018 · similarity 0.89)
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.
