How do humans sketch objects?

By Mathias Eitz, James Hays, Marc Alexa. 2012

1,037 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

EdgeConnect: Generative Image Inpainting with Adversarial Edge Learning (2019, background): “in the missing regions. Rather, we train an edge generator that hallucinates edges in these areas. Our approach of “lines ﬁrst, color next” is partly inspired by our understanding of how artists work [13]. “In line drawing, the lines not only delineate and deﬁne spaces and shapes; they also play a v”
Image-to-Image Translation with Conditional Adversarial Networks (2016, methodology): “• Sketch→photo: tests edges→photo models on human-drawn sketches from [17].”
Meta-attention for ViT-backed Continual Learning (2022, methodology): “To validate the superiority of the proposed method, extensive experiments, including benchmark comparison and ablation study, are conducted on a diverse set of image classification benchmarks (including ImageNet [8], CUB [47], Stanford Cars [24], FGVC-Aircraft [29], CIFAR-100 [25], Sketches [11], Wi”
Compacting, Picking and Growing for Unforgetting Continual Learning (2019, methodology): “The statistics are summarized in Table 4, where ImageNet [17] is the ﬁrst task, following by ﬁne-grained classiﬁcation tasks, CUBS [44], Stanford Cars [15] and Flowers [26], and ﬁnally WikiArt [39] and Sketch [8] that are artiﬁcial images drawing in various styles and objects.”
The Variational Gaussian Process (2015, methodology): “As a demonstration of the VGP’s complexity for learning representations, we also examine the Sketch data set (Eitz et al., 2012).”
Incremental Learning Through Deep Adaptation (2017, methodology): “following datasets: Caltech-256[33], CIFAR-10 [34], Daimler [35] (DPed), GTSR [36], Omniglot [37], Plankton imagery data [38] (Plnk), Human Sketch dataset [39] and SVHN [40].”
Deeper, Broader and Artier Domain Generalization (2017, neutral): “A more compelling motivation is domains where the total available images is fundamentally constrained, such as for particular styles of art [5, 29], and sketches [33, 6, 34, 26].”
DeLiGAN: Generative Adversarial Networks for Diverse and Limited Data (2017, methodology): “The TU-Berlin dataset [6], contains 20, 000 hand-drawn sketches evenly distributed among 250 object categories, which amounts to 80 images per category.”
Scribbler: Controlling Deep Image Synthesis with Sketch and Color (2016, background): “Their results are high quality and they are able to synthesize shoes and handbags from coarse sketches [10] even though their training data was simple image edges.”
SpotTune: Transfer Learning Through Adaptive Fine-Tuning (2018, methodology): “We compare our SpotTune method with other fine-tuning and regularization techniques on 5 public datasets, including three fine-grained classification benchmarks: CUBS [48], Stanford Cars [21] and Flowers [34], and two datasets with a large domain mismatch from ImageNet: Sketches [10] and WikiArt [41”
Zero-Shot Sketch-Image Hashing (2018, methodology): “To perform SBIR with binary codes under the novelly-deﬁned zero-shot cross-modal setting, the experiments of this work are taken on two large-scale sketch datasets, i.e ., Sketchy [52] and TU-Berlin [15], with extended images obtained from [39].”
Domain Adaptation by Mixture of Alignments of Second-or Higher-Order Scatter Tensors (2016, background): “rforms baselines (S+T) and (T) which demonstrates robustness of our approach. Heterogeneous setting on Pascal VOC2007-TU Berlin. Table 4 shows results on transfer from Pascal VOC2007 [8] to TU Berlin [7] (images-to-sketches transfer). As demonstrated in the table, our second-order approach (So) and ”
Deep Spatial-Semantic Attention for Fine-Grained Sketch-Based Image Retrieval (2017, background): “With the proliferation of touch-screen devices, a number of sketch-based computer vision problems have attracted increasing attention, including sketch recognition [47, 36, 3, 32], sketch-based image retrieval [46, 24, 10], sketch-based 3D model retrieval [39], and forensic sketch analysis [14, 28].”
StyleMeUp: Towards Style-Agnostic Sketch-Based Image Retrieval (2021, methodology): “For the second dataset, sketches are taken from the TU-Berlin Extension [13] which contains 250 object categories with 80 free-hand sketches per category.”
SalientShape: group saliency in image collections (2013, background): “Recent advances in human object sketch classifications [21] can potentially be used in con-”
SketchyGAN: Towards Diverse and Realistic Sketch to Image Synthesis (2018, methodology): “One of the most commonly used sketch dataset is the TU-Berlin dataset [10] which contains 20,000 human sketches spanning 250 categories.”