Research map / Label-efficient and robust learning / Recognition and long-tail
Binary Hashing with Semidefinite Relaxation and Augmented Lagrangian
Thanh-Toan Do, Anh-Dzung Doan, Duc-Thanh Nguyen, Ngai-Man Cheung
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 |
|---|---|---|---|
| infer codes with SDR instead of nonlinear optimization (NOPT) | two-step supervised hashing · CIFAR10, 8 bits | precision@2: 29.09 -> 30.57 | Improved |
| infer codes with SDR instead of NOPT | two-step supervised hashing · MNIST, 8 bits | precision@2: 81.14 -> 86.33 | Improved |
| infer codes with SDR instead of NOPT | two-step supervised hashing · SUN397, 8 bits | precision@2: 10.08 -> 12.11 | Improved |
| infer codes with AL instead of Graphcut | two-step supervised hashing · CIFAR10, 8 bits | precision@2: 28.50 -> 30.07 | Improved |
| use SDR instead of AL for binary inference | two-step hashing · CIFAR10, 32 bits | precision@2: 48.02 -> 48.43 | No clear effect |
| use SDR instead of AL for binary inference | two-step hashing · SUN397, 16 bits | precision@2: 57.20 -> 59.19 | Improved |
| use AL instead of SDR for binary inference | two-step hashing · SUN397, 24 bits | precision@2: 62.98 -> 63.14 | No clear effect |
| use SDR/AL codes instead of TSH | two-step supervised hashing · SUN397, 32 bits | precision@2: 59.30 -> 61.78 | Improved |
| use SDR instead of ITQ | unsupervised hashing · CIFAR10, 32 bits | precision@2: 15.36 -> 25.87 | Improved |
| use SDR instead of NMF | unsupervised hashing · CIFAR10, 32 bits classification | accuracy: 23.59 -> 29.65 | Improved |
| use AL with binary constraint in Lagrangian and tailored init instead of ALM | unsupervised hashing · MNIST, 32 bits classification | accuracy: 73.76 -> 81.13 | Improved |
Closest papers by meaning
- Learning to Hash with Binary Deep Neural Network (ECCV 2016 · similarity 0.88)
- Hashing with Binary Matrix Pursuit (ECCV 2018 · similarity 0.88)
- Affinity Preserving Quantization for Hashing: A Vector Quantization Approach to Learning Compact Binary Codes (AAAI 2016 · similarity 0.88)
- Relaxation-Free Deep Hashing via Policy Gradient (ECCV 2018 · similarity 0.87)
- Optimizing affinity-based binary hashing using auxiliary coordinates (NeurIPS 2016 · similarity 0.87)
- A Lower Bound of Hash Codes' Performance (NeurIPS 2022 · similarity 0.87)
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.
