Offline Sketch Parsing via Shapeness Estimation

Authors: Jie Wu, Changhu Wang, Liqing Zhang, Yong Rui

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

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments show the superiority of the proposed framework over stateof-the-art works on sketch parsing in both effectiveness and efficiency, even though they leveraged the temporal information of strokes. In this section, we evaluate the proposed shapeness estimation algorithm and the offline sketch parsing framework. The experiments were conducted on the FC Dataset [Lemaitre et al., 2013], a widely used benchmark dataset on sketch parsing.
Researcher Affiliation Collaboration Jie Wu,1 , Changhu Wang,2 Liqing Zhang,1 Yong Rui2 1Key Laboratory of Shanghai Education Commission for Intelligent Interaction and Cognitive Engineering, Department of Computer Science and Engineering, Shanghai Jiao Tong University 2Microsoft Research, Beijing, P. R. China
Pseudocode No The paper does not contain any structured pseudocode or algorithm blocks (e.g., clearly labeled 'Pseudocode' or 'Algorithm' sections).
Open Source Code No The paper does not provide any concrete access to source code (e.g., a specific repository link, explicit code release statement, or code in supplementary materials) for the methodology described.
Open Datasets Yes The experiments were conducted on the FC Dataset [Lemaitre et al., 2013], a widely used benchmark dataset on sketch parsing.
Dataset Splits No The paper states that "THdepth, THdist, and THaspect were set to 6, 0.04, and 7, which were learnt from the training set," implying the existence of a training set. However, it does not provide specific details on the dataset splits (e.g., percentages, sample counts, or explicit validation split information) for reproducibility.
Hardware Specification No The paper does not provide specific hardware details (e.g., exact GPU/CPU models, processor types, or memory amounts) used for running its experiments.
Software Dependencies No The paper refers to algorithms and techniques (e.g., "Graham Scan algorithm," "linear SVM," "1NN classifier") but does not provide specific version numbers for any software components or libraries used in their implementation.
Experiment Setup Yes In this work, THdepth, THdist, and THaspect were set to 6, 0.04, and 7, which were learnt from the training set. Finally, we only keep the stroke groups with scores higher than a predefined threshold, e.g., 0 in this work, for further shape recognition.