Video Temporal Grounding (VTG) involves Moment Retrieval (MR) and Highlight Detection (HD) based on textual queries. For this, most methods rely solely on final-layer features of frozen large pre-trained backbones, limiting their adaptability to new domains. While full fine-tuning is often impractical, parameter-efficient fine-tuning -- and particularly side-tuning (ST) -- has emerged as an effective alternative. However, prior ST approaches this problem from a frame-level refinement perspective, overlooking the inherent sparse nature of MR. To address this, we propose the Sparse-Dense Side-Tuner (SDST), the first anchor-free ST architecture for VTG. We also introduce the Reference-based Deformable Self-Attention, a novel mechanism that enhances the context modeling of the deformable attention -- a key limitation of existing anchor-free methods. Additionally, we present the first effective integration of InternVideo2 backbone into an ST framework, showing its profound implications in performance. Overall, our method significantly improves existing ST methods, achieving highly competitive or SOTA results on QVHighlights, TACoS, and Charades-STA, while reducing up to a 73% the parameter count w.r.t. the existing SOTA methods. The code is publicly accessible at https://github.com/davidpujol/SDST.

Temporal Action Localization (TAL) is a complex task that poses relevant challenges, particularly when attempting to generalize on new-unseen-domains in realworld applications. These scenarios, despite realistic, are often neglected in the literature, exposing these solutions to important performance degradation. In this work, we tackle this issue by introducing, for the first time, an approach for Unsupervised Domain Adaptation (UDA) in sparse TAL, which we refer to as Semantic Adversarial unsupervised Domain Adaptation (SADA). Our contributions are threefold: (1) we pioneer the development of a domain adaptation model that operates on realistic sparse action detection benchmarks; (2) we tackle the limitations of global-distribution alignment techniques by introducing a novel adversarial loss that is sensitive to local class distributions, ensuring finer-grained adaptation; and (3) we present a novel set of benchmarks based on EpicKitchens100 and CharadesEgo, that evaluate multiple domain shifts in a comprehensive manner. Our experiments indicate that SADA improves the adaptation across domains when compared to fully supervised state-of-the-art and alternative UDA methods, attaining a performance boost of up to 6.14% mAP. The code is publicly available at https://github.com/davidpujol/SADA.

Sign Language Translation (SLT) is a challenging task due to its cross-domain nature, involving the translation of visual-gestural language to text. Many previous methods employ an intermediate representation, i.e., gloss sequences, to facilitate SLT, thus transforming it into a two-stage task of sign language recognition (SLR) followed by sign language translation (SLT). However, the scarcity of gloss-annotated sign language data, combined with the information bottleneck in the mid-level gloss representation, has hindered the further development of the SLT task. To address this challenge, we propose a novel Gloss-Free SLT based on Visual-Language Pretraining (GFSLT-VLP), which improves SLT by inheriting language-oriented prior knowledge from pre-trained models, without any gloss annotation assistance. Our approach involves two stages: (i) integrating Contrastive Language-Image Pre-training (CLIP) with masked self-supervised learning to create pre-tasks that bridge the semantic gap between visual and textual representations and restore masked sentences, and (ii) constructing an end-to-end architecture with an encoder-decoder-like structure that inherits the parameters of the pre-trained Visual Encoder and Text Decoder from the first stage. The seamless combination of these novel designs forms a robust sign language representation and significantly improves gloss-free sign language translation. In particular, we have achieved unprecedented improvements in terms of BLEU-4 score on the PHOENIX14T dataset (≥+5) and the CSL-Daily dataset (≥+3) compared to state-of-the-art gloss-free SLT methods. Furthermore, our approach also achieves competitive results on the PHOENIX14T dataset when compared with most of the gloss-based methods1.

Transformer models have shown great success handling long-range interactions, making them a promising tool for modeling video. However, they lack inductive biases and scale quadratically with input length. These limitations are further exacerbated when dealing with the high dimensionality introduced by the temporal dimension. While there are surveys analyzing the advances of Transformers for vision, none focus on an in-depth analysis of video-specific designs. In this survey, we analyze the main contributions and trends of works leveraging Transformers to model video. Specifically, we delve into how videos are handled at the input level first. Then, we study the architectural changes made to deal with video more efficiently, reduce redundancy, re-introduce useful inductive biases, and capture long-term temporal dynamics. In addition, we provide an overview of different training regimes and explore effective self-supervised learning strategies for video. Finally, we conduct a performance comparison on the most common benchmark for Video Transformers (i.e., action classification), finding them to outperform 3D ConvNets even with less computational complexity.

This paper summarizes the ChaLearn Looking at People 2020 Challenge on Identity-preserved Human Detection (IPHD). For the purpose, we released a large novel dataset containing more than 112K pairs of spatiotemporally aligned depth and thermal frames (and 175K instances of humans) sampled from 780 sequences. The sequences contain hundreds of non-identifiable people appearing in a mix of in-the-wild and scripted scenarios recorded in public and private places. The competition was divided into three tracks depending on the modalities exploited for the detection: (1) depth, (2) thermal, and (3) depth-thermal fusion. Color was also captured but only used to facilitate the groundtruth annotation. Still the temporal synchronization of three sensory devices is challenging, so bad temporal matches across modalities can occur. Hence, the labels provided should considered “weak”, although test frames were carefully selected to minimize this effect and ensure the fairest comparison of the participants' results. Despite this added difficulty, the results got by the participants demonstrate current fully-supervised methods can deal with that and achieve outstanding detection performance when measured in terms of [email protected].

The interest in action and gesture recognition has grown considerably in the last years. In this paper, we present a survey on current deep learning methodologies for action and gesture recognition in image sequences. We introduce a taxonomy that summarizes important aspects of deep learning for approaching both tasks. We review the details of the proposed architectures, fusion strategies, main datasets, and competitions. We summarize and discuss the main works proposed so far with particular interest on how they treat the temporal dimension of data, discussing their main features and identify opportunities and challenges for future research.

This paper provides an overview of the Joint Contest on Multimedia Challenges Beyond Visual Analysis. We organized an academic competition that focused on four problems that require effective processing of multimodal information in order to be solved. Two tracks were devoted to gesture spotting and recognition from RGB-D video, two fundamental problems for human computer interaction. Another track was devoted to a second round of the first impressions challenge of which the goal was to develop methods to recognize personality traits from short video clips. For this second round we adopted a novel collaborative-competitive (i.e., coopetition) setting. The fourth track was dedicated to the problem of video recommendation for improving user experience. The challenge was open for about 45 days, and received outstanding participation: almost 200 participants registered to the contest, and 20 teams sent predictions in the final stage. The main goals of the challenge were fulfilled: the state of the art was advanced considerably in the four tracks, with novel solutions to the proposed problems (mostly relying on deep learning). However, further research is still required. The data of the four tracks will be available to allow researchers to keep making progress in the four tracks.