Face Clustering in Videos with Proportion Prior

Authors: Zhiqiang Tang, Yifan Zhang, Zechao Li, Hanqing Lu

IJCAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In experiments on a public data set from realworld videos, we observe improvements on clustering performance against state-of-the-art methods.
Researcher Affiliation Academia Zhiqiang Tang1, Yifan Zhang1 , Zechao Li2, Hanqing Lu1 1National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences 2School of Computer Science and Engineering, Nanjing University of Science and Technology zqtang2013@gmail.com, {yfzhang, luhq}@nlpr.ia.ac.cn, zechao.li@njust.edu.cn
Pseudocode Yes Algorithm 1 EM for Hidden Conditional Random Field
Open Source Code No The paper does not contain any explicit statement about releasing the source code for their methodology, nor does it provide a link to a code repository.
Open Datasets Yes We evaluate the performance of our method in the public face data set Big Bang Theory(BBT) given in [Bauml et al., 2013].
Dataset Splits No The paper describes sampling faces and computing track labels, but does not explicitly detail a training, validation, and test dataset split for model development or evaluation in the traditional sense, as it's a clustering task.
Hardware Specification No The paper does not provide any specific hardware details such as GPU models, CPU models, or memory specifications used for running the experiments.
Software Dependencies No The paper does not provide specific ancillary software details with version numbers (e.g., library names with versions).
Experiment Setup Yes We set λ1 = 10 4 in the experiments. λ2 is set to 10 4 in the experiments. ... the k-nearest neighbor graph is used and k = 10. ... The feature of each face is represented by a 240 dimensional Discrete Cosine Transform (DCT) vector. ... Then the Laplacian Eigenmaps reduces the feature dimension from 240 to the cluster number.