Zero-shot Learning with Deep Neural Networks for Object Recognition.

Zero-shot learning deals with the ability to recognize objects without any visual training sample. ... This chapter presents a review of the approaches based on deep neural networks to tackle the ZSL problem. ... The Synthesized Samples for Zero-Shot Learning [GDHG17] or SSZSL method similarly assumes that p(x|c) is Gaussian, estimates parameters (µ, Σ) for seen classes with techniques similar to GFZSL and aims ...

arXiv:2102.03137v1 fatcat:gyd4bxsbgrbivdwlwcqtca37w4

Open Access

inputs, and learning a shared attributes-based state-of-the art for zero-shot recognition on representation across modalities. ... zero-shot learning tasks. ...

doi:10.1109/cvpr.2016.13 dblp:conf/cvpr/ReedALS16 fatcat:nbnhrfhhvrbkjmncr74v3wcj5m

State-of-the-art methods for zero-shot visual recognition formulate learning as a joint embedding problem of images and side information. ... Our model achieves strong performance on zero-shot text-based image retrieval and significantly outperforms the attribute-based state-of-the-art for zero-shot classification on the Caltech UCSD Birds 200 ... CUB zero-shot recognition and retrieval In this section we describe the protocol and results for our zero-shot tasks. ...

arXiv:1605.05395v1 fatcat:csh33lsn6zbv7djlah5espgxxq

We propose a few-shot adaptation framework, which bridges zero-shot learning and supervised many-shot learning, for semantic indexing of image and video data. ... Few-shot adaptation provides robust parameter estimation with few training examples, by optimizing the parameters of zero-shot learning and supervised many-shot learning simultaneously. ... For example, support vector machines (SVMs) [8, 9] and deep neural networks [10] [11] [12] have been shown to be effective in video semantic indexing [2, 13] and object recognition [10] [11] [12 ...

arXiv:1807.07203v1 fatcat:xn2e3v44kvcalaub2ogpkwrkki

Few-shot adaptation provides robust parameter estimation with few training examples, by optimizing the parameters of zero-shot learning and supervised manyshot learning simultaneously. ... We propose a few-shot adaptation framework, which bridges zeroshot learning and supervised many-shot learning, for semantic indexing of image and video data. ... For example, support vector machines (SVMs) [8, 9] and deep neural networks [10] [11] [12] have been shown to be effective in video semantic indexing [2, 13] and object recognition [10] [11] [12 ...

doi:10.1145/3240508.3240592 dblp:conf/mm/InoueS18 fatcat:c2f7sik6cfegnc66ihbvw4bloe

In this work, we propose the application of a Siamese convolutional neural network to do one-shot recognition for such a task. ... With recent developments in deep learning and computer vision technologies, it has become possible to detect various features from the images with near-human accuracy. ... The approaches like few-shot learning and zero-shot learning can further reduce the data requirements for deep learning tasks [38, 30, 44, 34] . ...

arXiv:2005.05815v1 fatcat:swvmq2dvavalnokgf2tasjotbu

In the first feature learning phase, we finetune deep convolutional neural networks using hierarchical semantic structure among fine-grained classes to extract discriminative deep visual features. ... In addition, the features obtained by our feature learning model also yield significant gains when they are used by other zero-shot learning models, which shows the flexility of our model in zero-shot ... . / Zero-shot Fine-grained Classification by Deep Feature Learning with Semantics (See Fig. 2 ). ...

doi:10.1007/s11633-019-1177-8 fatcat:o6ve4m2o6zhodje5ckhjrvfoka

With the recent renaissance of deep convolution neural networks, encouraging breakthroughs have been achieved on the supervised recognition tasks, where each class has sufficient training data and fully ... One approach to scaling up the recognition is to develop models capable of recognizing unseen categories without any training instances, or zero-shot recognition/ learning. ... Yanwei Fu is supported by The Program for Professor of Special Appointment (Eastern Scholar) at Shanghai Institutions of Higher Learning. ...

arXiv:1710.04837v1 fatcat:u3mp6dgj2rgqrarjm4dcywegmy

We illustrate effectiveness of this semantic representation through experiments on zero-shot action/video classification and clustering. ... Our semantic fusion network combines three streams of information using a three-layer neural network: (i) frame-based low-level CNN features, (ii) object features from a state-of-the-art large-scale CNN ... Zero-shot Learning Baselines: We compare the following methods for largescale zero-shot recognition: 1. DAP-word. ...

doi:10.1109/cvpr.2016.339 dblp:conf/cvpr/WuFJS16 fatcat:6zpkaj64c5hx7g6hsttamlj5yq

The letter demonstrates that the object-prior learnt by a deep neural network while being trained for a denoising task can also be utilized for the phase retrieval, if the diffraction physics is effectively ... To alleviate this data requirement problem, this letter proposes a zero-shot learning method. ... For example, a neural formed zero-shot network), which combines the classical error- network trained to identify horses can be used to identify ze- reduction algorithm and deep learning for obtaining ...

arXiv:2106.04577v1 fatcat:o7j5pypztvdzjbmzydegpv6odm

Open Access

We evaluate our method on Animals with Attributes and Caltech-UCSD Birds 200-2011 dataset with a wide range of applications, including zero and few-shot image recognition and retrieval, from inductive ... Taking advantage of the recent success of unsupervised learning in deep neural networks, we propose an end-to-end learning framework that is able to extract more robust multi-modal representations across ... Figure 4 . 4 Fraction p of test images used for training ReViSE on transductive (a) zero-shot recognition (b) zero-shot retrieval for CUB dataset with att attributes. ...

arXiv:1703.05908v2 fatcat:bmvr3bbvavepbg7k7gwgks7gne

Multiple Versions

They significantly outperform the previous approaches that are not based on neural networks particularly for object category recognition. ... It has been recently reported that convolutional neural networks (CNNs) show good performances in many image recognition tasks. ... Transfer learning by deep neural networks The recent advances in the study of deep neural networks are initiated by the study of Hinton et al. [7] on unsupervised pretraining of deep networks. ...

doi:10.1007/978-3-319-16808-1_25 fatcat:uxo6jvvpdvaglf53v7zxwvahi4

Concepts such as Siamese Networks and triplet loss which are commonly used for one-shot learning will be examined. ... Concepts such as Siamese Networks and triplet loss which are commonly used for one-shot learning will be examined. ... In general, image representations are learned via a supervised metric-based approach with Siamese neural networks who's subnetwork's features are then used for one-shot learning so as to avoid having to ...

doi:10.1016/j.promfg.2020.01.025 fatcat:uwawbraov5hoplr3ymi2mrjkpa

Open Access

In this work, we propose the application of a Siamese convolutional neural network to do one-shot recognition for such a task. ... In this work, we propose the application of a Siamese convolutional neural network to do one-shot recognition for such a task. ... The approaches like few-shot learning and zero-shot learning can further reduce the data requirements for deep learning tasks [28, 29, 30, 31] . ...

doi:10.1016/j.promfg.2020.05.146 fatcat:zi3apgvtbzevbkuukjy54p7i6i

Open Access

Existing zero-shot learning (ZSL) models typically learn a projection function from a feature space to a semantic embedding space (e.g.~attribute space). ... We show that with this additional reconstruction constraint, the learned projection function from the seen classes is able to generalise better to the new unseen classes. ... Sparse feature learning for [26] D. Jayaraman and K. Grauman. Zero-shot recognition with deep belief networks. In Advances in neural information pro- unreliable attributes. ...

arXiv:1704.08345v1 fatcat:3tlzperd2nc3vmdmmigqavdonm

Zero-shot Learning with Deep Neural Networks for Object Recognition [article]

Preserved Fulltext

Learning Deep Representations of Fine-Grained Visual Descriptions

Preserved Fulltext

Learning Deep Representations of Fine-grained Visual Descriptions [article]

Preserved Fulltext

Few-Shot Adaptation for Multimedia Semantic Indexing [article]

Preserved Fulltext

Few-Shot Adaptation for Multimedia Semantic Indexing

Preserved Fulltext

One-Shot Recognition of Manufacturing Defects in Steel Surfaces [article]

Preserved Fulltext

Zero-shot Fine-grained Classification by Deep Feature Learning with Semantics

Preserved Fulltext

Recent Advances in Zero-shot Recognition [article]

Preserved Fulltext

Harnessing Object and Scene Semantics for Large-Scale Video Understanding

Preserved Fulltext

Phase retrieval with physics informed zero-shot learning [article]

Preserved Fulltext

Learning Robust Visual-Semantic Embeddings [article]

Preserved Fulltext

Other Versions

Understanding Convolutional Neural Networks in Terms of Category-Level Attributes [chapter]

Preserved Fulltext

One-Shot Learning for Custom Identification Tasks; A Review

Preserved Fulltext

One-Shot Recognition of Manufacturing Defects in Steel Surfaces

Preserved Fulltext

Semantic Autoencoder for Zero-Shot Learning [article]

Preserved Fulltext