Cross-domain Generative Learning for Fine-Grained Sketch-Based Image Retrieval

By Kaiyue Pang, Yi-Zhe Song, Tony Xiang, Timothy M. Hospedales. 2017

76 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

Doodle It Yourself: Class Incremental Learning by Drawing a Few Sketches (2022, background): “…triplet-ranking based Siamese networks for joint embedding space learning, which was further reinforced by attention [53], cross-domain translation [35], re-inforcement learning based on-the-ﬂy retrieval [6], semi-supervised retrieval [2], style-agnostic retrieval [46], etc. Sketch for Vision…”
Fine-Grained Image Analysis With Deep Learning: A Survey (2021, methodology): “domain gap in FG-SBIR was tackled via a discriminativegenerative hybrid model [182] using cross-domain translation.”
More Photos are All You Need: Semi-Supervised Learning for Fine-Grained Sketch Based Image Retrieval (2021, background): “Subsequent works have aimed at improving this via attention mechanisms with higher order retrieval loss [39], joint discriminative-generative learning with cross-modal image generation [27], text tags [38], and cross-modal hierarchical co-attention [33].”
StyleMeUp: Towards Style-Agnostic Sketch-Based Image Retrieval (2021, methodology): “This was further enhanced via attention based techniques with advanced higher order retrieval loss [49], hybrid generativediscriminative cross-domain image generation [36], textual tags [48] and employing mixed modal jigsaw solving for a better pre-training strategy [37].”
Sketching without Worrying: Noise-Tolerant Sketch-Based Image Retrieval (2022, unknown): “This was further augmented via attention with higher-order retrieval loss [46], cross-domain image generation [34], text tags [45], etc.”
Sketch Less for More: On-the-Fly Fine-Grained Sketch-Based Image Retrieval (2020, unknown): “A number of deep learning approaches subsequently emerged [49, 41, 30, 29].”
Cross-Modal Hierarchical Modelling for Fine-Grained Sketch Based Image Retrieval (2020, methodology): “Starting from the study of deformable-part models [21], recent methods have learned to extract comparable features from heterogeneous domains [29, 39, 50].”
Sketch3T: Test-Time Training for Zero-Shot SBIR (2022, background): “Attention mechanisms along with higher-order losses [54], hybrid generative-discriminative cross-domain image generation [39], textual tags [53] and mixed-modal jigsaw solving based pre-training strategy [40], enhanced it further.”
Image-to-image translation for cross-domain disentanglement (2018, background): “In cross-domain retrieval [4, 24, 34], query and database are from different domains.”
Partially Does It: Towards Scene-Level FG-SBIR with Partial Input (2022, methodology): “This framework was subsequently enhanced through a hybrid generative-discriminative cross-domain image generation [53], providing an attention based mechanism with advanced higher order retrieval loss [64], utilising textual tags [63], or pre-training strategy [5, 54].”
Deep Learning for Free-Hand Sketch: A Survey (2020, background): “Sharing some similar motivations with [165], generative adversarial network based SBIR models also have been studied.”
Semantically Tied Paired Cycle Consistency for Zero-Shot Sketch-Based Image Retrieval (2019, background): “[33] introduced a discriminative-generative hybrid model for preserving all the domain invariant information useful for reducing the domain gap between sketch and image.”
Generalising Fine-Grained Sketch-Based Image Retrieval (2019, methodology): “∗Equal Contribution Recent FG-SBIR methods [24, 36, 28, 22] address this issue by learning a deep network embedding of sketch and photo that makes them directly comparable.”
Fine-Grained Image Retrieval: the Text/Sketch Input Dilemma (2017, methodology): “Deep neural network (DNN) provides an end-to-end solution for ﬁne-grained SBIR [17, 19, 21, 22, 33], and are proven to be superior to shallow methods based on hand-crafted features [12, 13].”