Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].
Dual-Reference Face Retrieval
Authors: BingZhang Hu, Feng Zheng, Ling Shao
AAAI 2018 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | The experiments show promising results, outperforming hierarchical methods. |
| Researcher Affiliation | Collaboration | 1School of Computing Sciences, University of East Anglia, Norwich, UK 2Department of Electrical and Computer Engineering, University of Pittsburgh, Pittsburgh, USA 3JD Artificial Intelligence Research (JDAIR), Beijing, China |
| Pseudocode | No | The paper describes methods and a network architecture, but does not include explicitly labeled pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not contain an explicit statement about releasing source code or a link to a code repository for the described methodology. |
| Open Datasets | Yes | In the experiment, we evaluate our DRFR on three face recognition and age estimation datasets: Cross-Age Celebrity Dataset(CACD) (Chen, Chen, and Hsu 2014), FGNet (Lanitis and Cootes 2002), and MORPH (Ricanek and Tesafaye 2006). |
| Dataset Splits | No | The paper states, "we take 60% data as training data and the remaining for the test," but does not explicitly mention a separate validation set or its split percentage. |
| Hardware Specification | No | The paper does not specify any hardware components (e.g., GPU models, CPU types, memory) used for running the experiments. |
| Software Dependencies | No | The paper does not list any specific software dependencies or their version numbers required for reproducibility. |
| Experiment Setup | Yes | For the hyper-parameters, we set the ε in Eq. 11 as 5 to calculate the similarity matrix set S, and the embeddings size on the joint manifold is set as 128. [...] we employed an online quartet selection protocol which is inspired by (Chen et al. 2017). During training, the images of an entire mini batch are firstly propagated forward to extract the embeddings with the current model, then those quartets which violate the average margin in this mini batch will be selected to train the network. |