Improving Generalization via Scalable Neighborhood Component Analysis

Wu, Zhirong; Efros, Alexei A.; Yu, Stella X.

doi:10.1007/978-3-030-01234-2_42

Cited by 116 publications

(129 citation statements)

References 39 publications

Supporting

Mentioning

128

Contrasting

Order By: Relevance

“…Metric Learning. Metric learning approaches [35,22] have achieved remarkable performance on different vision tasks, such as image retrieval [70,72,71] and face recognition [65,69,60]. Such tasks usually involve open world recognition, since classes during testing might be disjoint from the ones in the training set.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

SegSort: Segmentation by Discriminative Sorting of Segments

Hwang

Shi

et al. 2019

2019 IEEE/CVF International Conference on Computer Vision (ICCV)

View full text Add to dashboard Cite

Almost all existing deep learning approaches for semantic segmentation tackle this task as a pixel-wise classification problem. Yet humans understand a scene not in terms of pixels, but by decomposing it into perceptual groups and structures that are the basic building blocks of recognition. This motivates us to propose an end-to-end pixelwise metric learning approach that mimics this process. In our approach, the optimal visual representation determines the right segmentation within individual images and associates segments with the same semantic classes across images. The core visual learning problem is therefore to maximize the similarity within segments and minimize the similarity between segments. Given a model trained this way, inference is performed consistently by extracting pixel-wise embeddings and clustering, with the semantic label determined by the majority vote of its nearest neighbors from an annotated set. As a result, we present the SegSort, as a first attempt using deep learning for unsupervised semantic segmentation, achieving 76% performance of its supervised counterpart. When supervision is available, Seg-Sort shows consistent improvements over conventional approaches based on pixel-wise softmax training. Additionally, our approach produces more precise boundaries and consistent region predictions. The proposed SegSort further produces an interpretable result, as each choice of label can be easily understood from the retrieved nearest segments.

show abstract

Section: Related Workmentioning

confidence: 99%

“…When supervision is available (i.e., supervised semantic segmentation), we segment each image with the spherical K-Means clustering and train the network following the same optimization, but incorporated with Neighborhood Components Analysis criterion [22,71] for semantic labels.…”

Section: Introductionmentioning

confidence: 99%

SegSort: Segmentation by Discriminative Sorting of Segments

Hwang

Shi

et al. 2019

2019 IEEE/CVF International Conference on Computer Vision (ICCV)

View full text Add to dashboard Cite

show abstract

“…These metric learning strategies have been widely used in image retrieval [53], face recognition [54,55] and person reidentification [56]. More recently, Wu et al [43] presented a feature embedding method based on neighborhood component analysis. These works show that combining deep models with proper objectives is effective in learning the similarities.…”

Section: B Deep Metric Learningmentioning

confidence: 99%

“…The second branch are metric based approaches [2,[31][32][33][40][41][42][43]. Metric learning based methods learn a set of project functions (embedding functions) and metrics to measure the similarity between the query and samples images and classify them in a feed-forward manner.…”

Section: Metric Learning For Few-shot Learningmentioning

confidence: 99%

Revisiting metric learning for few-shot image classification

et al. 2020

Neurocomputing

View full text Add to dashboard Cite

The goal of few-shot learning is to recognize new visual concepts with just a few amount of labeled samples in each class. Recent effective metric-based few-shot approaches employ neural networks to learn a feature similarity comparison between query and support examples. However, the importance of feature embedding, i.e., exploring the relationship among training samples, is neglected. In this work, we present a simple yet powerful baseline for few-shot classification by emphasizing the importance of feature embedding. Specifically, we revisit the classical triplet network from deep metric learning, and extend it into a deep K-tuplet network for few-shot learning, utilizing the relationship among the input samples to learn a general representation learning via episode-training. Once trained, our network is able to extract discriminative features for unseen novel categories and can be seamlessly incorporated with a non-linear distance metric function to facilitate the few-shot classification. Our result on the miniImageNet benchmark outperforms other metric-based few-shot classification methods. More importantly, when evaluated on completely different datasets (Caltech-101, CUB-200, Stanford Dogs and Cars) using the model trained with miniImageNet, our method significantly outperforms prior methods, demonstrating its superior capability to generalize to unseen classes.

show abstract

“…For example, the relation network has a 65.32% accuracy for 5-way 5-shot setting on miniImageNet [39]. gives a 61.1%;[2] has 66.6%;[21] obtains 71.07% with a larger network…”

mentioning

confidence: 99%

Finding Task-Relevant Features for Few-Shot Learning by Category Traversal

Eigen

Dodge

et al. 2019

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

318

159

View full text Add to dashboard Cite

Few-shot learning is an important area of research. Conceptually, humans are readily able to understand new concepts given just a few examples, while in more pragmatic terms, limited-example training situations are common in practice. Recent effective approaches to few-shot learning employ a metric-learning framework to learn a feature similarity comparison between a query (test) example, and the few support (training) examples. However, these approaches treat each support class independently from one another, never looking at the entire task as a whole. Because of this, they are constrained to use a single set of features for all possible test-time tasks, which hinders the ability to distinguish the most relevant dimensions for the task at hand. In this work, we introduce a Category Traversal Module that can be inserted as a plug-and-play module into most metric-learning based few-shot learners. This component traverses across the entire support set at once, identifying task-relevant features based on both intra-class commonality and inter-class uniqueness in the feature space. Incorporating our module improves performance considerably (5%-10% relative) over baseline systems on both mini-ImageNet and tieredImageNet benchmarks, with overall performance competitive with recent state-of-the-art systems.

show abstract

Improving Generalization via Scalable Neighborhood Component Analysis

Cited by 116 publications

References 39 publications

SegSort: Segmentation by Discriminative Sorting of Segments

SegSort: Segmentation by Discriminative Sorting of Segments

Revisiting metric learning for few-shot image classification

Finding Task-Relevant Features for Few-Shot Learning by Category Traversal

Contact Info

Product

Resources

About