Fine-Grained Visual Comparisons with Local Learning

By Aron Yu, Kristen Grauman. 2014

584 citations recorded by the source corpus.

Subfields: Artificial Intelligence, 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 Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

GAN Inversion: A Survey (2021, methodology): “128 Z [38], [44] [39], [45] first using the term inversion”
Generative Visual Manipulation on the Natural Image Manifold (2016, methodology): “Datasets: We experiment with multiple photo collections from various sources as follows: “shoes” dataset [37], which has 50K shoes collected from Zappos.”
Unpaired Image-to-Image Translation Using Cycle-Consistent Adversarial Networks (2017, neutral): “Figure8shows some example results on other paired datasets used in “pix2pix” [22], such as architectural labels$photos from the CMP Facade Database [40], and edges$shoes from the UT Zappos50K dataset [60]. The image quality of our results is close to those produced by the fully supervised pix2pix wh”
Multimodal Unsupervised Image-to-Image Translation (2018, methodology): “[ We use the datasets provided by Isola et al. [6], Yu et al 80], and Zhu et al. [81], which contain images of shoes and handbags with edge maps generated by HED [82].”
Learning to Discover Cross-Domain Relations with Generative Adversarial Networks (2017, methodology): “We validated that our DiscoGAN performs very well on this type of image-to-image translation task and generate realistic photos of handbags (Zhu et al., 2016) and shoes (Yu & Grauman, 2014).”
Recent advances in convolutional neural networks (2015, background): “There are already some finegrained image datasets (such as Flower [98], Birds [99], [100], Dogs [101], Cars [102] and Shoes [103]).”
Piggyback GAN: Efficient Lifelong Learning for Image Conditioned Generation (2021, methodology): “We first demonstrate the effectiveness of Piggyback GAN on 4 tasks of paired image-conditioned generation, which are all image-to-image translations, on challenging domains and datasets with large variation across different modalities [9,14,35,39].”
Lifelong GAN: Continual Learning for Conditional Image Generation (2019, background): “The first task is image-to-image translation of edges → shoes photos [37, 35].”
Toward Multimodal Image-to-Image Translation (2017, methodology): “Datasets We test our method on several image-to-image translation problems from prior work, including edges→ photos [48, 54], Google maps→ satellite [20], labels→ images [5], and outdoor night→ day images [25].”
ALICE: Towards Understanding Adversarial Learning for Joint Distribution Matching (2017, background): “(ii) Edge-to-Shoe [21]: x domain consists of shoe photos and z domain consists of edge images, on which we report extensive quantitative comparisons.”
The iNaturalist Challenge 2017 Dataset (2017, background): “In contrast, mass-produced, man-made object categories are typically identical and only differ in terms of pose, lighting, color, but not necessarily in their underlying object shape or appearance [47, 6, 48].”
Generative Compression (2017, methodology): “(a) Randomly sampled images from the CelebA dataset (left) [4], Zappos50k dataset (middle) [5] and Outdoor Places dataset (right) [6].”
JointGAN: Multi-Domain Joint Distribution Learning with Generative Adversarial Nets (2018, methodology): “Datasets We present results on ﬁve datasets: edges ↔ shoes (Yu & Grauman, 2014), edges ↔ handbags (Zhu et al., 2016), Google maps ↔ aerial photos (Isola et al., 2017), labels ↔ facades (Tyle ˇ cek & Šára, 2013) and labels ↔ cityscapes (Cordts et al., 2016).”
Perceptual Adversarial Networks for Image-to-Image Transformation (2017, methodology): “The original data is from [46] and [56], and the HED edge detector [57] was used to extract edges.”
A Decade Survey of Content Based Image Retrieval Using Deep Learning (2020, background): “The shoes apparel dataset, namely UT-ZAP50K [50], consists of roughly 50K images.”
Interactive Sketch & Fill: Multiclass Sketch-to-Image Translation (2019, methodology): “We first compare our 2 step approach for interactive image generation on existing datasets such as the UTZappos Shoes dataset [57] and CelebA-HQ [26].”
Fine-Grained Image Classification by Exploring Bipartite-Graph Labels (2015, background): “Yu and Grauman [60] fixed this issue by learning local functions that tailor the comparisons to neighborhood statistics of the data.”
Fashion IQ: A New Dataset Towards Retrieving Images by Natural Language Feedback (2021, background): “Many fashion datasets have been proposed over the past few years, covering different applications such as fashionability and style prediction [50, 28, 22, 51], fashion image generation [46], product search and recommendation [25, 72, 19, 41, 63], fashion apparel pixelwise segmentation [27, 74, 69], ”