A performance evaluation of gradient field HOG descriptor for sketch based image retrieval

By Rui Hu, John Philip Collomosse. Computer Vision and Image Understanding, 2013

403 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/f4268671d089b90cb52fe10bc77499cceb66d00e

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

Deep Learning for Free-Hand Sketch: A Survey (2020, background): “The generation across sketch and 3D model is also an interesting cross-modal research topic.”
Zero-Shot Sketch-Image Hashing (2018, background): “SBIR, including ﬁne-grained SBIR, learns shared representations to speciﬁcally mitigate the expressional gap between hand-crafted sketches and real images [16, 17, 24, 25, 47, 48, 49, 50, 52, 56, 58, 61, 68, 69, 76, 77], while the efﬁciency issue is not considered.”
Sketch-a-Net: A Deep Neural Network that Beats Humans (2016, methodology): “…including sketch recognition (Eitz et al, 2012; Schneider and Tuytelaars, 2014; Yu et al, 2015), sketch-based image retrieval (Eitz et al, 2011; Hu and Collomosse, 2013), sketchbased shape (Zou et al, 2015) or 3D model retrieval (Wang et al, 2015), and forensic sketch analysis (Klare et al,…”
Cross-Modal Hierarchical Modelling for Fine-Grained Sketch Based Image Retrieval (2020, background): “2 Related Work Category-level SBIR: Most category-level SBIR approaches fall into two categories: (i) handcrafted descriptors [32, 41] which constructs global [31] or local [14] photo-sketch joint representations.”
3D deep shape descriptor (2015, methodology): “The bag-of-features (BOF) method is used to extract a frequency histogram of geometric words for shape retrieval in previous works [13, 14, 22].”
Sketch3T: Test-Time Training for Zero-Shot SBIR (2022, methodology): “Aiming to retrieve photos of the same category, categorylevel SBIR [12, 49] began with using handcrafted descriptors [60] like SIFT [31], Gradient Field HOG [22], Histogram of Edge Local Orientations [44] or Learned Key Shapes [45], for constructing local [22] or global [42] joint photo-sketch repre”
StyleMeUp: Towards Style-Agnostic Sketch-Based Image Retrieval (2021, methodology): “Early approaches deploy handcrafted descriptors [52] such as SIFT [31], Gradient Field HOG [19], Histogram of Edge Local Orientations [42] or Learned Key Shapes [43], for constructing local [19] or global [40] joint photo-sketch representations.”
Sketch-Guided Object Localization in Natural Images (2020, background): “Despite the problem of sketch-guided object localization has a relation to the well-studied sketch-based image retrieval (SBIR) problem [5,14,20,25,28,30,38,40], these are distinctively different in terms of objectives.”
A Zero-Shot Framework for Sketch-based Image Retrieval (2018, methodology): “Hand-crafted feature based models include the gradient field HOG descriptor proposed by Hu and Collomose [21], the histogram of edge orientations (HELO) proposed by Saavendra [22], the learned key shapes (LKS) proposed by Saavendra et.al [23] which are used in Bag of Visual Words (BoVW) framework as”
SketchyGAN: Towards Diverse and Realistic Sketch to Image Synthesis (2018, background): “There exist numerous works on sketch-based image retrieval [12, 13, 22, 3, 4, 53, 24, 23, 27, 52, 38, 54, 35].”
Deep Sketch Hashing: Fast Free-Hand Sketch-Based Image Retrieval (2017, methodology): “Being able to better handle the domain gap, deep methods typically achieve higher performance than hand-crafted ones for both category-level [13, 19, 49, 73, 45, 50, 12] and finegrained [52, 66, 27] SBIR tasks.”
Semantically Tied Paired Cycle Consistency for Zero-Shot Sketch-Based Image Retrieval (2019, background): “Hand-crafted methods mostly work by extracting the edge map from natural image and then matching them with sketch using a Bag-of-Words model on top of some specifically designed SBIR features, viz., gradient field HOG [20], histogram of oriented edges [40], learned key shapes [41] etc.”
Sketchformer: Transformer-Based Representation for Sketched Structure (2020, background): “sketch based image retrieval (SBIR) techniques explored region [12], graph [13], edge-let [14], and sparse gradient features [15, 16] building upon the success of dictionary learning based models (e.”
Cross-Domain Adversarial Feature Learning for Sketch Re-identification (2018, background): “There are many works [7, 8, 12, 31] focusing on category-level sketch classification and sketch-to-photo retrieval.”
Semantic-Aware Knowledge Preservation for Zero-Shot Sketch-Based Image Retrieval (2019, methodology): “Early works employed hand-crafted features to represent the sketches and matched them with the edge maps extracted from the photo images using different variants of the Bag-Of-Words model [32, 13, 7, 14, 6].”
The sketchy database (2016, methodology): “In fact, this strategy is consistent with common sketch-based image retrieval benchmarks [Hu and Collomosse 2013] where retrieval results are correct as long as they are the same category.”
Sketch Me That Shoe (2016, methodology): “Therefore, with few exceptions [22, 11], most existing SBIR datasets were created by combining overlapping categories of sketches and photos, which means only categorylevel SBIR is possible.”
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].”
Generalising Fine-Grained Sketch-Based Image Retrieval (2019, background): “Fine-grained SBIR Most earlier SBIR studies [5, 10, 4, 14, 38, 7, 20] focus on category-level cross-domain matching.”