Chuang Gan

DevNet: A Deep Event Network for multimedia event detection and evidence recounting

In this paper, we focus on complex event detection in internet videos while also providing the key evidences of the detection results. Convolutional Neural Networks (CNNs) have achieved promising performance in image classification and action recognition tasks. However, it remains an open problem how to use CNNs for video event detection and recounting, mainly due to the complexity and diversity of video events. In this work, we propose a flexible deep CNN infrastructure, namely Deep Event Network (DevNet), that simultaneously detects pre-defined events and provides key spatial-temporal evidences. Taking key frames of videos as input, we first detect the event of interest at the video level by aggregating the CNN features of the key frames. The pieces of evidences which recount the detection results, are also automatically localized, both temporally and spatially. The challenge is that we only have video level labels, while the key evidences usually take place at the frame levels. Based on the intrinsic property of CNNs, we first generate a spatial-temporal saliency map by back passing through DevNet, which then can be used to find the key frames which are most indicative to the event, as well as to localize the specific spatial position, usually an object, in the frame of the highly indicative area. Experiments on the large scale TRECVID 2014 MEDTest dataset demonstrate the promising performance of our method, both for event detection and evidence recounting.

Semantic Compositional Networks for Visual Captioning

A Semantic Compositional Network (SCN) is developed for image captioning, in which semantic concepts (i.e., tags) are detected from the image, and the probability of each tag is used to compose the parameters in a long short-term memory (LSTM) network. The SCN extends each weight matrix of the LSTM to an ensemble of tag-dependent weight matrices. The degree to which each member of the ensemble is used to generate an image caption is tied to the image-dependent probability of the corresponding tag. In addition to captioning images, we also extend the SCN to generate captions for video clips. We qualitatively analyze semantic composition in SCNs, and quantitatively evaluate the algorithm on three benchmark datasets: COCO, Flickr30k, and Youtube2Text. Experimental results show that the proposed method significantly outperforms prior state-of-the-art approaches, across multiple evaluation metrics.

Learning Attributes Equals Multi-Source Domain Generalization

Attributes possess appealing properties and benefit many computer vision problems, such as object recognition, learning with humans in the loop, and image retrieval. Whereas the existing work mainly pursues utilizing attributes for various computer vision problems, we contend that the most basic problem-how to accurately and robustly detect attributes from images-has been left under explored. Especially, the existing work rarely explicitly tackles the need that attribute detectors should generalize well across different categories, including those previously unseen. Noting that this is analogous to the objective of multi-source domain generalization, if we treat each category as a domain, we provide a novel perspective to attribute detection and propose to gear the techniques in multi-source domain generalization for the purpose of learning cross-category generalizable attribute detectors. We validate our understanding and approach with extensive experiments on four challenging datasets and three different problems.

You Lead, We Exceed: Labor-Free Video Concept Learning by Jointly Exploiting Web Videos and Images

Video concept learning often requires a large set oftraining samples. In practice, however, acquiring noise-free training labels with sufficient positive examples is very expensive. A plausible solution for training data collection is by sampling from the vast quantities of images and videos on the Web. Such a solution is motivated by the assumption that the retrieved images or videos are highly correlated with the query. Still, a number ofchallenges remain. First, Web videos are often untrimmed. Thus, only parts of the videos are relevant to the query. Second, the retrieved Web images are always highly relevant to the issued query. However, thoughtlessly utilizing the images in the video domain may even hurt the performance due to the well-known semantic drift and domain gap problems. As a result, a valid question is how Web images and videos interact for video concept learning. In this paper, we propose a Lead-Exceed Neural Network (LENN), which reinforces the training on Web images and videos in a curriculum manner. Specifically, the training proceeds by inputting frames of Web videos to obtain a network. The Web images are then filtered by the learnt network and the selected images are additionally fed into the network to enhance the architecture and further trim the videos. In addition, Long Short-Term Memory (LSTM) can be applied on the trimmed videos to explore temporal information. Encouraging results are reported on UCFIOl, TRECVID 2013 and 2014 MEDTest in the context ofboth action recognition and event detection. Without using human annotated exemplars, our proposed LENN can achieve 74.4% accuracy on UCFIOI dataset.

Webly-Supervised Video Recognition by Mutually Voting for Relevant Web Images and Web Video Frames

Video recognition usually requires a large amount of training samples, which are expensive to be collected. An alternative and cheap solution is to draw from the large-scale images and videos from the Web. With modern search engines, the top ranked images or videos are usually highly correlated to the query, implying the potential to harvest the labeling-free Web images and videos for video recognition. However, there are two key difficulties that prevent us from using the Web data directly. First, they are typically noisy and may be from a completely different domain from that of users’ interest (e.g. cartoons). Second, Web videos are usually untrimmed and very lengthy, where some query-relevant frames are often hidden in between the irrelevant ones. A question thus naturally arises: to what extent can such noisy Web images and videos be utilized for labeling-free video recognition? In this paper, we propose a novel approach to mutually voting for relevant Web images and video frames, where two forces are balanced, i.e. aggressive matching and passive video frame selection. We validate our approach on three large-scale video recognition datasets.

StyleNet: Generating Attractive Visual Captions with Styles

We propose a novel framework named StyleNet to address the task of generating attractive captions for images and videos with different styles. To this end, we devise a novel model component, named factored LSTM, which automatically distills the style factors in the monolingual text corpus. Then at runtime, we can explicitly control the style in the caption generation process so as to produce attractive visual captions with the desired style. Our approach achieves this goal by leveraging two sets of data: 1) factual image/video-caption paired data, and 2) stylized monolingual text data (e.g., romantic and humorous sentences). We show experimentally that StyleNet outperforms existing approaches for generating visual captions with different styles, measured in both automatic and human evaluation metrics on the newly collected FlickrStyle10K image caption dataset, which contains 10K Flickr images with corresponding humorous and romantic captions.

Recognizing an Action Using Its Name: A Knowledge-Based Approach

Existing action recognition algorithms require a set of positive exemplars to train a classifier for each action. However, the amount of action classes is very large and the users’ queries vary dramatically. It is impractical to pre-define all possible action classes beforehand. To address this issue, we propose to perform action recognition with no positive exemplars, which is often known as the zero-shot learning. Current zero-shot learning paradigms usually train a series of attribute classifiers and then recognize the target actions based on the attribute representation. To ensure the maximum coverage of ad-hoc action classes, the attribute-based approaches require large numbers of reliable and accurate attribute classifiers, which are often unavailable in the real world. In this paper, we propose an approach that merely takes an action name as the input to recognize the action of interest without any pre-trained attribute classifiers and positive exemplars. Given an action name, we first build an analogy pool according to an external ontology, and each action in the analogy pool is related to the target action at different levels. The correlation information inferred from the external ontology may be noisy. We then propose an algorithm, namely adaptive multi-model rank-preserving mapping (AMRM), to train a classifier for action recognition, which is able to evaluate the relatedness of each video in the analogy pool adaptively. As multiple mapping models are employed, our algorithm has better capability to bridge the gap between visual features and the semantic information inferred from the ontology. Extensive experiments demonstrate that our method achieves the promising performance for action recognition only using action names, while no attributes and positive exemplars are available.

A Multisource Domain Generalization Approach to Visual Attribute Detection

Attributes possess appealing properties and benefit many computer vision problems, such as object recognition, learning with humans in the loop, and image retrieval. Whereas the existing work mainly pursues utilizing attributes for various computer vision problems, we contend that the most basic problem—how to accurately and robustly detect attributes from images—has been left underexplored. Especially, the existing work rarely explicitly tackles the need that attribute detectors should generalize well across different categories, including those previously unseen. Noting that this is analogous to the objective of multisource domain generalization, if we treat each category as a domain, we provide a novel perspective to attribute detection and propose to gear the techniques in multisource domain generalization for the purpose of learning cross-category generalizable attribute detectors. We validate our understanding and approach with extensive experiments on four challenging datasets and two different problems.