Bayesian Multiple Target Localization
Authors: Purnima Rajan, Weidong Han, Raphael Sznitman, Peter Frazier, Bruno Jedynak
ICML 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We present an empirical evaluation of this policy on simulated data for the problem of detecting multiple instances of the same object in an image. Finally, we present experiments on localizing multiple faces simultaneously on real images. |
| Researcher Affiliation | Academia | Purnima Rajan PURNIMA@CS.JHU.EDU Department of Computer Science, Johns Hopkins University Weidong Han WHAN@PRINCETON.EDU Department of Operations Research and Financial Engineering, Princeton University Raphael Sznitman RAPHAEL.SZNITMAN@ARTORG.UNIBE.CH ARTORG Center, University of Bern Peter I. Frazier PF98@CORNELL.EDU School of Operations Research and Information Engineering, Cornell University Bruno M. Jedynak BRUNO.JEDYNAK@JHU.EDU Department of Applied Mathematics & Statistics, Johns Hopkins University |
| Pseudocode | Yes | Algorithm 1 Posterior Rank (PR) Algorithm... Algorithm 2 Iterated Posterior Rank (IPR) Algorithm |
| Open Source Code | No | The paper does not provide any explicit statements about releasing source code or links to code repositories for the described methodology. |
| Open Datasets | Yes | We begin by training an extremely efficient but poorlyperforming face classifier... using 4000 faces and 5 million background samples of size 30 30... We evaluated our poor-classifier at multiple scales on 35 images from the MIT+CMU face dataset. |
| Dataset Splits | No | The paper mentions training data and simulated data, but does not specify any training/validation/test splits, either by percentage or absolute counts, for reproducibility. |
| Hardware Specification | No | The paper does not specify any hardware details such as GPU models, CPU types, or memory used for running the experiments. |
| Software Dependencies | No | The paper mentions training a 'Boosted classifier' and cites a paper (Ali et al., 2012) for its description, but does not provide specific software names with version numbers for any libraries or tools used in the experiments. |
| Experiment Setup | Yes | We use 100 random assignments for the locations of the object instances for each k and each image size in the simulation... We use the additive model presented in (19) and we choose Wn to be independent, Normally distributed random variable with standard deviation σ... for two levels of noise σ = 0.5, 1. We begin by training an extremely efficient but poorlyperforming face classifier... with 50 stumps... using 4000 faces and 5 million background samples of size 30 30. |