Research map / Label-efficient and robust learning / Recognition and long-tail
Multi-scale Classification using Localized Spatial Depth
Subhajit Dutta, Soham Sarkar, Anil K. Ghosh
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 |
|---|---|---|---|
| use localized spatial depth (LSPD) instead of plain SPD features | depth-based GAM classifier · Simulated E4 (d=5) | misclassification rate (%): 10.07 -> 2.58 | Improved |
| use localized spatial depth (LSPD) instead of plain SPD features | depth-based GAM classifier · Parkinson's | misclassification rate (%): 15.44 -> 11.35 | Improved |
| use localized spatial depth (LSPD) instead of plain SPD features | depth-based GAM classifier · Growth | misclassification rate (%): 14.64 -> 5.10 | Improved |
| use localized spatial depth (LSPD) instead of plain SPD features | depth-based GAM classifier · Simulated E2 (d=100) | misclassification rate (%): 2.90 -> 3.18 | Hurt |
| use localized spatial depth (LSPD) instead of plain SPD features | depth-based GAM classifier · Vehicle | misclassification rate (%): 16.35 -> 17.15 | Hurt |
| use localized spatial depth (LSPD) instead of plain SPD features | depth-based GAM classifier · Colon | misclassification rate (%): 19.98 -> 20.51 | No clear effect |
| use spatial depth features instead of k-NN | k-NN classifier · Simulated E1 (d=100) | misclassification rate (%): 44.57 -> 1.60 | Improved |
| use LSPD features instead of KDE | KDE classifier · Simulated E3 (d=100) | misclassification rate (%): 49.98 -> 0.00 | Improved |
| use SPD features instead of depth-depth (DD) plot classifier | DD classifier · Simulated E2 (d=5) | misclassification rate (%): 26.68 -> 9.26 | Improved |
| use LSPD features instead of DD classifier | DD classifier · Simulated E4 (d=5) | misclassification rate (%): 33.04 -> 2.58 | Improved |
| use LSPD instead of KDE | KDE classifier · Parkinson's | misclassification rate (%): 11.01 -> 11.35 | No clear effect |
| use multiscale aggregation over localization levels h instead of a single h | LSPD classifier · Simulated E3 and E4 | misclassification rate (%) | Improved |
Closest papers by meaning
- Non-Local Spatial Propagation Network for Depth Completion (ECCV 2020 · similarity 0.82)
- Divide&Classify: Fine-Grained Classification for City-Wide Visual Geo-Localization (ICCV 2023 · similarity 0.82)
- Probabilistic Framework for the Characterization of Surfaces and Edges in Range Images, with Application to Edge Detection (TPAMI 2017 · similarity 0.82)
- Local Feature Selection for Data Classification (TPAMI 2016 · similarity 0.82)
- LocalBins: Improving Depth Estimation by Learning Local Distributions (ECCV 2022 · similarity 0.81)
- Scaling Image Geo-Localization to Continent Level (NeurIPS 2025 · similarity 0.81)
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.
