Transductive Zero-Shot Action Recognition by Word-Vector Embedding

Zero-shot learning (ZSL) aims to classify unseen samples based on the relationship between the learned visual features and semantic features. Traditional ZSL methods typically capture the underlying multimodal data structures by learning an embedding function between the visual space and the semantic space with the Euclidean metric. However, these models suffer from the hubness problem and domain bias problem, which leads to unsatisfactory performance, especially in the generalized ZSL (GZSL) task. To tackle such a problem, we formulate a discriminative cross-aligned variational autoencoder (DCA-VAE) for ZSL. The proposed model effectively utilizes a modified cross-modal-alignment variational autoencoder (VAE) to transform both visual features and semantic features obtained by the discriminative cosine metric into latent features. The key to our method is that we collect principal discriminative information from visual and semantic features to construct latent features which contain the discriminative multimodal information associated with unseen samples. Finally, the proposed model DCA-VAE is validated on six benchmarks including the large dataset ImageNet, and several experimental results demonstrate the superiority of DCA-VAE over most existing embedding or generative ZSL models on the standard ZSL and the more realistic GZSL tasks.

Zero-shot learning (ZSL) typically explores a shared semantic space in order to recognize novel categories in the absence of any labeled training data. However, the traditional ZSL methods always suffer from serious domain shift problem in human action recognition. This is because: 1) existing ZSL methods are specifically designed for object recognition from static images, which do not capture the temporal dynamics of video sequences, and poor performances are always generated if those methods are directly applied to zero-shot action recognition; 2) these methods always blindly project the target data into a shared space using a semantic mapping obtained by the source data without any adaptation, in which the underlying structures of target data are ignored; and 3) severe inter-class variations exist in various action categories. The traditional ZSL methods do not take relationships across different categories into consideration. In this paper, we propose a novel aligned dynamic-preserving embedding (ADPE) model for zero-shot action recognition in a transductive setting. In our model, an adaptive embedding of target videos is learned, exploring the distributions of both the source and target data. An aligned regularization is further proposed to couple the centers of target semantic representations with their corresponding label prototypes in order to preserve the relationships across different categories. Most significantly, during our embedding, the temporal dynamics of video sequences are simultaneously preserved via exploiting the temporal consistency of video sequences and capturing the temporal evolution of successive segments of actions. Our model can effectively overcome the domain shift problem in zero-shot action recognition. The experiments on Olympic sports, HMDB51, and UCF101 datasets demonstrate the effectiveness of our model.

Most zero-shot learning (ZSL) methods aim to learn a mapping from visual feature space to semantic feature space or from both visual and semantic feature spaces to a common joint space and align them. However, in these methods the visual and semantic information are not utilized sufficiently and the useless information is not excluded. Moreover, there exists a strong bias problem that the instances from unseen classes always tend to be predicted as some seen classes in most ZSL methods. In this paper, combining the advantages of generative adversarial networks (GANs), a method based on bidirectional projections between the visual and semantic feature spaces is proposed. GANs are used to perform bidirectional generations and alignments between the visual and semantic features. In addition, cycle mapping structure ensures that the important information are kept in the alignments. Furthermore, in order to better solve the bias problem, pseudo-labels are generated for unseen instances and the model is adjusted along with them iteratively. We conduct extensive experiments at traditional ZSL and generalized ZSL settings, respectively. Experiment results confirm that our method achieves the state-of-the-art performances on the popular datasets AWA2, aPY and SUN.

Zero-shot learning (ZSL), a type of structured multioutput learning, has attracted much attention due to its requirement of no training data for target classes. Conventional ZSL methods usually project visual features into semantic space and assign labels by finding their nearest prototypes. However, this type of nearest neighbor search (NNS)-based method often suffers from great performance degradation because of the nonuniform variances between different categories. In this article, we propose a probabilistic framework by taking covariance into account to deal with the above-mentioned problem. In this framework, we define a new latent space, which has two characteristics. The first is that the features in this space should gather within the classes and scatter between the classes, which is implemented by triplet learning; the second is that the prototypes of unseen classes are synthesized with nonnegative coefficients, which are generated by nonnegative matrix factorization (NMF) of relations between the seen classes and the unseen classes in attribute space. During training, the learned parameters are the projection model for triplet network and the nonnegative coefficients between the unseen classes and the seen classes. In the testing phase, visual features are projected into latent space and assigned with the labels that have the maximum probability among unseen classes for classic ZSL or within all classes for generalized ZSL. Extensive experiments are conducted on four popular data sets, and the results show that the proposed method can outperform the state-of-the-art methods in most circumstances.

Recognition of human activities is an essential field in computer vision. The most human activity consists of the interaction between humans and objects. Many successful works have been done on human-object interaction (HOI) recognition and achieved acceptable results in recent years. Still, they are fully supervised and need to train labeled data for all HOIs. Due to the enormous space of human-object interactions, listing and providing the training data for all possible categories is costly and impractical. We propose an approach for scaling human-object interaction recognition in video data through the zero-shot learning technique to solve this problem. Our method recognizes a verb and an object from the video and makes an HOI class. Recognition of the verbs and objects instead of HOIs allows identifying a new combination of verbs and objects. So, a new HOI class can be identified, which is not seen by the recognizer system. We introduce a neural network architecture that can understand and represent the video data. The proposed system learns verbs and objects from available training data at the training phase and can identify the verb-object pairs in a video at test time. So, the system can identify the HOI class with different combinations of objects and verbs. Also, we propose to use lateral information for combining the verbs and the objects to make valid verb-object pairs. It helps to prevent the detection of rare and probably wrong HOIs. The lateral information comes from word embedding techniques. Furthermore, we propose a new feature aggregation method for aggregating extracted high-level features from video frames before feeding them to the classifier. We illustrate that this feature aggregation method is more effective for actions that include multiple subactions. We evaluated our system by recently introduced Charades challengeable dataset, which has lots of HOI categories in videos. We show that our proposed system can detect unseen HOI classes in addition to the acceptable recognition of seen types. Therefore, the number of classes identifiable by the system is greater than the number of classes used for training.

Zero shot learning (ZSL) provides a solution to recognising unseen classes without class labelled data for model learning. Most ZSL methods aim to learn a mapping from a visual feature space to a semantic embedding space, e.g. attribute or word vector spaces. The use of word vector space is particularly attractive as compared to attribute, it offers vast auxiliary classes with free parts embedding without human annotation. However, using the word vector embedding often provides weaker discriminative power than manually labelled attributes of the auxiliary classes. This is compounded further in zero-shot action recognition due to richer content variations among action classes. In this work we propose to explore a broader semantic contextual information in the text domain to enrich the word vector representation of action classes. We show through extensive experiments that this method improves significantly the performance of a number of existing word vector embedding ZSL methods. Moreover, it also outperforms attribute embedding ZSL with human annotation.

Unlike conventional zero-shot learning (CZSL) which only focuses on the recognition of unseen classes by using the classifier trained on seen classes and semantic embeddings, generalized zero-shot learning (GZSL) aims at recognizing both the seen and unseen classes, so it is more challenging due to the extreme training imbalance. Recently, some feature generation methods introduce metric learning to enhance the discriminability of visual features. Although these methods achieve good results, they focus only on metric learning in the visual feature space to enhance features and ignore the association between the feature space and the semantic space. Since the GZSL method uses semantics as prior knowledge to migrate visual knowledge to unseen classes, the consistency between visual space and semantic space is critical. To this end, we propose relational metric learning which can relate the metrics in the two spaces and make the distribution of the two spaces more consistent. Based on the generation method and relational metric learning, we proposed a novel GZSL method, termed VS-Boost, which can effectively boost the association between vision and semantics. The experimental results demonstrate that our method is effective and achieves significant gains on five benchmark datasets compared with the state-of-the-art methods.

Zero-shot learning (ZSL) for visual recognition aims to accurately recognize the objects of unseen classes through mapping the visual feature to an embedding space spanned by class semantic information. However, the semantic gap across visual features and their underlying semantics is still a big obstacle in ZSL. Conventional ZSL methods construct that the mapping typically focus on the original visual features that are independent of the ZSL tasks, thus degrading the prediction performance. In this paper, we propose an effective method to uncover an appropriate latent representation of data for the purpose of zero-shot classification. Specifically, we formulate a novel framework to jointly learn the latent subspace and cross-modal embedding to link visual features with their semantic representations. The proposed framework combines feature learning and semantics prediction, such that the learned data representation is more discriminative to predict the semantic vectors, hence improving the overall classification performance. To learn a robust latent subspace, we explicitly avoid the information loss by ensuring the reconstruction ability of the obtained data representation. An efficient algorithm is designed to solve the proposed optimization problem. To fully exploit the intrinsic geometric structure of data, we develop a manifold regularization strategy to refine the learned semantic representations, leading to further improvements of the classification performance. To validate the effectiveness of the proposed approach, extensive experiments are conducted on three ZSL benchmarks and encouraging results are achieved compared with the state-of-the-art ZSL methods.

Zero-shot learning (ZSL) has recently attracted increasing attention in visual tasks like action recognition. We propose a spatiotemporal visual-semantic embedding network (STVSEM) for zero-shot action recognition. First, given the fact that two-stream architecture based action recognition algorithms have achieved excellent results in recent years, the module is assembled to our designed network by simultaneously using the spatial features (e.g., RGB appearance) and optical flow in time domain as visual features to significantly improve the visual expression capability. Then, in order to slightly alleviate the problem of semantic loss that typically occurs in the case of using embedding-based ZSL methods, an autoencoder is introduced to get a better semantic representation and complement semantic relationship information for unseen classes by seen classes. Last but not least, a joint embedding mechanism that explores and exploits the relationships of the visual data and semantic information in an intermediate space is employed to ameliorate the gap between vision and semantics. The experimental results on Charades and UCF101 datasets indicate that the proposed method outperforms the state-of-the-art methods in accuracy, which further demonstrates the effectiveness of our method.

We investigate learning feature-to-feature translator networks by alternating back-propagation as a general-purpose solution to zero-shot learning (ZSL) problems. It is a generative model-based ZSL framework. In contrast to models based on generative adversarial networks (GAN) or variational autoencoders (VAE) that require auxiliary networks to assist the training, our model consists of a single conditional generator that maps class-level semantic features and Gaussian white noise vectors accounting for instance-level latent factors to visual features, and is trained by maximum likelihood estimation. The training process is a simple yet effective alternating back-propagation process that iterates the following two steps: (i) the inferential back-propagation to infer the latent noise vector of each observed example, and (ii) the learning back-propagation to update the model parameters. We show that, with slight modifications, our model is capable of learning from incomplete visual features for ZSL. We conduct extensive comparisons with existing generative ZSL methods on five benchmarks, demonstrating the superiority of our method in not only ZSL performance but also convergence speed and computational cost. Specifically, our model outperforms the existing state-of-the-art methods by a remarkable margin up to 3.1% and 4.0% in ZSL and generalized ZSL settings, respectively.

Zero-shot learning (ZSL) aims to recognize image instances of unseen classes solely based on the semantic descriptions of the unseen classes. In this field, Generalized Zero-Shot Learning (GZSL) is a challenging problem in which the images of both seen and unseen classes are mixed in the testing phase of learning. Existing methods formulate GZSL as a semantic-visual correspondence problem and apply generative models such as Generative Adversarial Networks and Variational Autoencoders to solve the problem. However, these methods suffer from the bias problem since the images of unseen classes are often misclassified into seen classes. In this work, a novel model named the Dual Projective model for Zero-Shot Learning (DPZSL) is proposed using text descriptions. In order to alleviate the bias problem, we leverage two autoencoders to project the visual and semantic features into a latent space and evaluate the embeddings by a visual-semantic correspondence loss function. An additional novel classifier is also introduced to ensure the discriminability of the embedded features. Our method focuses on a more challenging inductive ZSL setting in which only the labeled data from seen classes are used in the training phase. The experimental results, obtained from two popular datasets—Caltech-UCSD Birds-200-2011 (CUB) and North America Birds (NAB)—show that the proposed DPZSL model significantly outperforms both the inductive ZSL and GZSL settings. Particularly in the GZSL setting, our model yields an improvement up to 15.2% in comparison with state-of-the-art CANZSL on datasets CUB and NAB with two splittings.

Open zero-shot learning via asymmetric VAE with dissimilarity space

Most zero-shot learning (ZSL) methods based on generative adversarial networks (GANs) utilize random noise and semantic descriptions to synthesize image features of unseen classes, which alleviates the problem of training data imbalance between seen and unseen classes. However, these methods usually only learn the distributions of seen classes in the training stage, ignoring the unseen ones. Due to the different distributions of seen and unseen samples, i.e., image features, these methods cannot generate unseen features of sufficient quality, so the performances are also limited, especially for the generalized zero-shot learning (GZSL) setting. In this article, we propose a general transductive method based on GANs, called GT-GAN, which can improve the quality of generated unseen image features and therefore benefit the classification. A new loss function is introduced to make the relative positions between each unseen image and its <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$k$ </tex-math></inline-formula> nearest neighbors in the feature space as consistent as possible with their relative positions in the semantic space; this loss function may be easily applied in most existing GAN-based models. Experimental results on five benchmark datasets show a significant improvement in accuracy compared with that of original models, especially in the GZSL setting.

Pre-trained vision-language models, exemplified by CLIP, have exhibited promising zero-shot capabilities across various downstream tasks. Trained on image-text pairs, CLIP is naturally extendable to video-based action recognition, due to the similarity between processing images and video frames. To leverage this inherent synergy, numerous efforts have been directed towards adapting CLIP for action recognition tasks in videos. However, the specific methodologies for this adaptation remain an open question. Common approaches include prompt tuning and fine-tuning with or without extra model components on video-based action recognition tasks. Nonetheless, such adaptations may compromise the generalizability of the original CLIP framework and also necessitate the acquisition of new training data, thereby undermining its inherent zero-shot capabilities. In this study, we propose zero-shot action recognition (ZAR) by adapting the CLIP pre-trained model without the need for additional training datasets. Our approach leverages the entropy minimization technique, utilizing the current video test sample and augmenting it with varying frame rates. We encourage the model to make consistent decisions, and use this consistency to dynamically update a prompt learner during inference. Experimental results demonstrate that our ZAR method achieves state-of-the-art zero-shot performance on the Kinetics-600, HMDB51, and UCF101 datasets.

Zero-shot Learning (ZSL), which aims to predict for those classes that have never appeared in the training data, has arisen hot research interests. The key of implementing ZSL is to leverage the prior knowledge of classes which builds the semantic relationship between classes and enables the transfer of the learned models (e.g., features) from training classes (i.e., seen classes) to unseen classes. However, the priors adopted by the existing methods are relatively limited with incomplete semantics. In this paper, we explore richer and more competitive prior knowledge to model the inter-class relationship for ZSL via ontology-based knowledge representation and semantic embedding. Meanwhile, to address the data imbalance between seen classes and unseen classes, we developed a generative ZSL framework with Generative Adversarial Networks (GANs). Our main findings include: (i) an ontology-enhanced ZSL framework that can be applied to different domains, such as image classification (IMGC) and knowledge graph completion (KGC); (ii) a comprehensive evaluation with multiple zero-shot datasets from different domains, where our method often achieves better performance than the state-of-the-art models. In particular, on four representative ZSL baselines of IMGC, the ontology-based class semantics outperform the previous priors e.g., the word embeddings of classes by an average of 12.4 accuracy points in the standard ZSL across two example datasets (see Figure 4).