Efstratios Gavves

Modeling video evolution for action recognition

In this paper we present a method to capture video-wide temporal information for action recognition. We postulate that a function capable of ordering the frames of a video temporally (based on the appearance) captures well the evolution of the appearance within the video. We learn such ranking functions per video via a ranking machine and use the parameters of these as a new video representation. The proposed method is easy to interpret and implement, fast to compute and effective in recognizing a wide variety of actions. We perform a large number of evaluations on datasets for generic action recognition (Hollywood2 and HMDB51), fine-grained actions (MPII- cooking activities) and gestures (Chalearn). Results show that the proposed method brings an absolute improvement of 7-10%, while being compatible with and complementary to further improvements in appearance and local motion based methods.

Fine-Grained Categorization by Alignments

The aim of this paper is fine-grained categorization without human interaction. Different from prior work, which relies on detectors for specific object parts, we propose to localize distinctive details by roughly aligning the objects using just the overall shape, since implicit to fine-grained categorization is the existence of a super-class shape shared among all classes. The alignments are then used to transfer part annotations from training images to test images (supervised alignment), or to blindly yet consistently segment the object in a number of regions (unsupervised alignment). We furthermore argue that in the distinction of fine grained sub-categories, classification-oriented encodings like Fisher vectors are better suited for describing localized information than popular matching oriented features like HOG. We evaluate the method on the CU-2011 Birds and Stanford Dogs fine-grained datasets, outperforming the state-of-the-art.

Dynamic Image Networks for Action Recognition

We introduce the concept of dynamic image, a novel compact representation of videos useful for video analysis especially when convolutional neural networks (CNNs) are used. The dynamic image is based on the rank pooling concept and is obtained through the parameters of a ranking machine that encodes the temporal evolution of the frames of the video. Dynamic images are obtained by directly applying rank pooling on the raw image pixels of a video producing a single RGB image per video. This idea is simple but powerful as it enables the use of existing CNN models directly on video data with fine-tuning. We present an efficient and effective approximate rank pooling operator, speeding it up orders of magnitude compared to rank pooling. Our new approximate rank pooling CNN layer allows us to generalize dynamic images to dynamic feature maps and we demonstrate the power of our new representations on standard benchmarks in action recognition achieving state-of-the-art performance.

Siamese Instance Search for Tracking

In this paper we present a tracker, which is radically different from state-of-the-art trackers: we apply no model updating, no occlusion detection, no combination of trackers, no geometric matching, and still deliver state-of-the-art tracking performance, as demonstrated on the popular online tracking benchmark (OTB) and six very challenging YouTube videos. The presented tracker simply matches the initial patch of the target in the first frame with candidates in a new frame and returns the most similar patch by a learned matching function. The strength of the matching function comes from being extensively trained generically, i.e., without any data of the target, using a Siamese deep neural network, which we design for tracking. Once learned, the matching function is used as is, without any adapting, to track previously unseen targets. It turns out that the learned matching function is so powerful that a simple tracker built upon it, coined Siamese INstance search Tracker, SINT, which only uses the original observation of the target from the first frame, suffices to reach state-of-the-art performance. Further, we show the proposed tracker even allows for target re-identification after the target was absent for a complete video shot.

Guiding the Long-Short Term Memory Model for Image Caption Generation

In this work we focus on the problem of image caption generation. We propose an extension of the long short term memory (LSTM) model, which we coin gLSTM for short. In particular, we add semantic information extracted from the image as extra input to each unit of the LSTM block, with the aim of guiding the model towards solutions that are more tightly coupled to the image content. Additionally, we explore different length normalization strategies for beam search to avoid bias towards short sentences. On various benchmark datasets such as Flickr8K, Flickr30K and MS COCO, we obtain results that are on par with or better than the current state-of-the-art.

COSTA: Co-Occurrence Statistics for Zero-Shot Classification

In this paper we aim for zero-shot classification, that is visual recognition of an unseen class by using knowledge transfer from known classes. Our main contribution is COSTA, which exploits co-occurrences of visual concepts in images for knowledge transfer. These inter-dependencies arise naturally between concepts, and are easy to obtain from existing annotations or web-search hit counts. We estimate a classifier for a new label, as a weighted combination of related classes, using the co-occurrences to define the weight. We propose various metrics to leverage these co-occurrences, and a regression model for learning a weight for each related class. We also show that our zero-shot classifiers can serve as priors for few-shot learning. Experiments on three multi-labeled datasets reveal that our proposed zero-shot methods, are approaching and occasionally outperforming fully supervised SVMs. We conclude that co-occurrence statistics suffice for zero-shot classification.

Rank Pooling for Action Recognition

We propose a function-based temporal pooling method that captures the latent structure of the video sequence data - e.g., how frame-level features evolve over time in a video. We show how the parameters of a function that has been fit to the video data can serve as a robust new video representation. As a specific example, we learn a pooling function via ranking machines. By learning to rank the frame-level features of a video in chronological order, we obtain a new representation that captures the video-wide temporal dynamics of a video, suitable for action recognition. Other than ranking functions, we explore different parametric models that could also explain the temporal changes in videos. The proposed functional pooling methods, and rank pooling in particular, is easy to interpret and implement, fast to compute and effective in recognizing a wide variety of actions. We evaluate our method on various benchmarks for generic action, fine-grained action and gesture recognition. Results show that rank pooling brings an absolute improvement of 7-10 average pooling baseline. At the same time, rank pooling is compatible with and complementary to several appearance and local motion based methods and features, such as improved trajectories and deep learning features.

Locality in Generic Instance Search from One Example

This paper aims for generic instance search from a single example. Where the state-of-the-art relies on global image representation for the search, we proceed by including locality at all steps of the method. As the first novelty, we consider many boxes per database image as candidate targets to search locally in the picture using an efficient point-indexed representation. The same representation allows, as the second novelty, the application of very large vocabularies in the powerful Fisher vector and VLAD to search locally in the feature space. As the third novelty we propose an exponential similarity function to further emphasize locality in the feature space. Locality is advantageous in instance search as it will rest on the matching unique details. We demonstrate a substantial increase in generic instance search performance from one example on three standard datasets with buildings, logos, and scenes from 0.443 to 0.620 in mAP.

Local Alignments for Fine-Grained Categorization

The aim of this paper is fine-grained categorization without human interaction. Different from prior work, which relies on detectors for specific object parts, we propose to localize distinctive details by roughly aligning the objects using just the overall shape. Then, one may proceed to the classification by examining the corresponding regions of the alignments. More specifically, the alignments are used to transfer part annotations from training images to unseen images (supervised alignment), or to blindly yet consistently segment the object in a number of regions (unsupervised alignment). We further argue that for the distinction of sub-classes, distribution-based features like color Fisher vectors are better suited for describing localized appearance of fine-grained categories than popular matching oriented shape-sensitive features, like HOG. They allow capturing the subtle local differences between subclasses, while at the same time being robust to misalignments between distinctive details. We evaluate the local alignments on the CUB-2011 and on the Stanford Dogs datasets, composed of 200 and 120, visually very hard to distinguish bird and dog species. In our experiments we study and show the benefit of the color Fisher vector parameterization, the influence of the alignment partitioning, and the significance of object segmentation on fine-grained categorization. We, furthermore, show that by using object detectors as voters to generate object confidence saliency maps, we arrive at fully unsupervised, yet highly accurate fine-grained categorization. The proposed local alignments set a new state-of-the-art on both the fine-grained birds and dogs datasets, even without any human intervention. What is more, the local alignments reveal what appearance details are most decisive per fine-grained object category.

VideoLSTM convolves, attends and flows for action recognition

We present a new architecture for end-to-end sequence learning of actions in video, we call VideoLSTM. Rather than adapting the video to the peculiarities of established recurrent or convolutional architectures, we adapt the architecture to fit the requirements of the video medium. Starting from the soft-Attention LSTM, VideoLSTM makes three novel contributions. First, video has a spatial layout. To exploit the spatial correlation we hardwire convolutions in the soft-Attention LSTM architecture. Second, motion not only informs us about the action content, but also guides better the attention towards the relevant spatio-temporal locations. We introduce motion-based attention. And finally, we demonstrate how the attention from VideoLSTM can be used for action localization by relying on just the action class label. Experiments and comparisons on challenging datasets for action classification and localization support our claims.