Andrew Zisserman

Jean-Baptiste Alayrac, Jeff Donahue, Pauline Luc et al. · deepmind

Building models that can be rapidly adapted to novel tasks using only a handful of annotated examples is an open challenge for multimodal machine learning research. We introduce Flamingo, a family of Visual Language Models (VLM) with this ability. We propose key architectural innovations to: (i) bridge powerful pretrained vision-only and language-only models, (ii) handle sequences of arbitrarily interleaved visual and textual data, and (iii) seamlessly ingest images or videos as inputs. Thanks to their flexibility, Flamingo models can be trained on large-scale multimodal web corpora containing arbitrarily interleaved text and images, which is key to endow them with in-context few-shot learning capabilities. We perform a thorough evaluation of our models, exploring and measuring their ability to rapidly adapt to a variety of image and video tasks. These include open-ended tasks such as visual question-answering, where the model is prompted with a question which it has to answer; captioning tasks, which evaluate the ability to describe a scene or an event; and close-ended tasks such as multiple-choice visual question-answering. For tasks lying anywhere on this spectrum, a single Flamingo model can achieve a new state of the art with few-shot learning, simply by prompting the model with task-specific examples. On numerous benchmarks, Flamingo outperforms models fine-tuned on thousands of times more task-specific data.

Olivier J. Hénaff, Skanda Koppula, Evan Shelhamer et al. · deepmind

The promise of self-supervised learning (SSL) is to leverage large amounts of unlabeled data to solve complex tasks. While there has been excellent progress with simple, image-level learning, recent methods have shown the advantage of including knowledge of image structure. However, by introducing hand-crafted image segmentations to define regions of interest, or specialized augmentation strategies, these methods sacrifice the simplicity and generality that makes SSL so powerful. Instead, we propose a self-supervised learning paradigm that discovers this image structure by itself. Our method, Odin, couples object discovery and representation networks to discover meaningful image segmentations without any supervision. The resulting learning paradigm is simpler, less brittle, and more general, and achieves state-of-the-art transfer learning results for object detection and instance segmentation on COCO, and semantic segmentation on PASCAL and Cityscapes, while strongly surpassing supervised pre-training for video segmentation on DAVIS.

Adrià Recasens, Jason Lin, Joāo Carreira et al. · deepmind

Attention-based models are appealing for multimodal processing because inputs from multiple modalities can be concatenated and fed to a single backbone network - thus requiring very little fusion engineering. The resulting representations are however fully entangled throughout the network, which may not always be desirable: in learning, contrastive audio-visual self-supervised learning requires independent audio and visual features to operate, otherwise learning collapses; in inference, evaluation of audio-visual models should be possible on benchmarks having just audio or just video. In this paper, we introduce Zorro, a technique that uses masks to control how inputs from each modality are routed inside Transformers, keeping some parts of the representation modality-pure. We apply this technique to three popular transformer-based architectures (ViT, Swin and HiP) and show that with contrastive pre-training Zorro achieves state-of-the-art results on most relevant benchmarks for multimodal tasks (AudioSet and VGGSound). Furthermore, the resulting models are able to perform unimodal inference on both video and audio benchmarks such as Kinetics-400 or ESC-50.

Ragav Sachdeva, Andrew Zisserman

The goal of this paper is to detect what has changed, if anything, between two "in the wild" images of the same 3D scene acquired from different camera positions and at different temporal instances. The open-set nature of this problem, occlusions/dis-occlusions due to the shift in viewpoint, and the lack of suitable training datasets, presents substantial challenges in devising a solution. To address this problem, we contribute a change detection model that is trained entirely on synthetic data and is class-agnostic, yet it is performant out-of-the-box on real world images without requiring fine-tuning. Our solution entails a "register and difference" approach that leverages self-supervised frozen embeddings and feature differences, which allows the model to generalise to a wide variety of scenes and domains. The model is able to operate directly on two RGB images, without requiring access to ground truth camera intrinsics, extrinsics, depth maps, point clouds, or additional before-after images. Finally, we collect and release a new evaluation dataset consisting of real-world image pairs with human-annotated differences and demonstrate the efficacy of our method. The code, datasets and pre-trained model can be found at: https://github.com/ragavsachdeva/CYWS-3D

Liliane Momeni, Hannah Bull, K R Prajwal et al. · cambridge, oxford

Recently, sign language researchers have turned to sign language interpreted TV broadcasts, comprising (i) a video of continuous signing and (ii) subtitles corresponding to the audio content, as a readily available and large-scale source of training data. One key challenge in the usability of such data is the lack of sign annotations. Previous work exploiting such weakly-aligned data only found sparse correspondences between keywords in the subtitle and individual signs. In this work, we propose a simple, scalable framework to vastly increase the density of automatic annotations. Our contributions are the following: (1) we significantly improve previous annotation methods by making use of synonyms and subtitle-signing alignment; (2) we show the value of pseudo-labelling from a sign recognition model as a way of sign spotting; (3) we propose a novel approach for increasing our annotations of known and unknown classes based on in-domain exemplars; (4) on the BOBSL BSL sign language corpus, we increase the number of confident automatic annotations from 670K to 5M. We make these annotations publicly available to support the sign language research community.

Gül Varol, Liliane Momeni, Samuel Albanie et al. · cambridge

The focus of this work is $\textit{sign spotting}$ - given a video of an isolated sign, our task is to identify $\textit{whether}$ and $\textit{where}$ it has been signed in a continuous, co-articulated sign language video. To achieve this sign spotting task, we train a model using multiple types of available supervision by: (1) $\textit{watching}$ existing footage which is sparsely labelled using mouthing cues; (2) $\textit{reading}$ associated subtitles (readily available translations of the signed content) which provide additional $\textit{weak-supervision}$; (3) $\textit{looking up}$ words (for which no co-articulated labelled examples are available) in visual sign language dictionaries to enable novel sign spotting. These three tasks are integrated into a unified learning framework using the principles of Noise Contrastive Estimation and Multiple Instance Learning. We validate the effectiveness of our approach on low-shot sign spotting benchmarks. In addition, we contribute a machine-readable British Sign Language (BSL) dictionary dataset of isolated signs, BSLDict, to facilitate study of this task. The dataset, models and code are available at our project page.

K R Prajwal, Hannah Bull, Liliane Momeni et al. · cambridge, oxford

The goal of this work is to detect and recognize sequences of letters signed using fingerspelling in British Sign Language (BSL). Previous fingerspelling recognition methods have not focused on BSL, which has a very different signing alphabet (e.g., two-handed instead of one-handed) to American Sign Language (ASL). They also use manual annotations for training. In contrast to previous methods, our method only uses weak annotations from subtitles for training. We localize potential instances of fingerspelling using a simple feature similarity method, then automatically annotate these instances by querying subtitle words and searching for corresponding mouthing cues from the signer. We propose a Transformer architecture adapted to this task, with a multiple-hypothesis CTC loss function to learn from alternative annotation possibilities. We employ a multi-stage training approach, where we make use of an initial version of our trained model to extend and enhance our training data before re-training again to achieve better performance. Through extensive evaluations, we verify our method for automatic annotation and our model architecture. Moreover, we provide a human expert annotated test set of 5K video clips for evaluating BSL fingerspelling recognition methods to support sign language research.

Carl Doersch, Yi Yang, Mel Vecerik et al.

We present a novel model for Tracking Any Point (TAP) that effectively tracks any queried point on any physical surface throughout a video sequence. Our approach employs two stages: (1) a matching stage, which independently locates a suitable candidate point match for the query point on every other frame, and (2) a refinement stage, which updates both the trajectory and query features based on local correlations. The resulting model surpasses all baseline methods by a significant margin on the TAP-Vid benchmark, as demonstrated by an approximate 20% absolute average Jaccard (AJ) improvement on DAVIS. Our model facilitates fast inference on long and high-resolution video sequences. On a modern GPU, our implementation has the capacity to track points faster than real-time, and can be flexibly extended to higher-resolution videos. Given the high-quality trajectories extracted from a large dataset, we demonstrate a proof-of-concept diffusion model which generates trajectories from static images, enabling plausible animations. Visualizations, source code, and pretrained models can be found on our project webpage.

Carl Doersch, Ankush Gupta, Larisa Markeeva et al.

Generic motion understanding from video involves not only tracking objects, but also perceiving how their surfaces deform and move. This information is useful to make inferences about 3D shape, physical properties and object interactions. While the problem of tracking arbitrary physical points on surfaces over longer video clips has received some attention, no dataset or benchmark for evaluation existed, until now. In this paper, we first formalize the problem, naming it tracking any point (TAP). We introduce a companion benchmark, TAP-Vid, which is composed of both real-world videos with accurate human annotations of point tracks, and synthetic videos with perfect ground-truth point tracks. Central to the construction of our benchmark is a novel semi-automatic crowdsourced pipeline which uses optical flow estimates to compensate for easier, short-term motion like camera shake, allowing annotators to focus on harder sections of video. We validate our pipeline on synthetic data and propose a simple end-to-end point tracking model TAP-Net, showing that it outperforms all prior methods on our benchmark when trained on synthetic data.

Chang Liu, Yujie Zhong, Andrew Zisserman et al.

In this paper, we consider the problem of generalised visual object counting, with the goal of developing a computational model for counting the number of objects from arbitrary semantic categories, using arbitrary number of "exemplars", i.e. zero-shot or few-shot counting. To this end, we make the following four contributions: (1) We introduce a novel transformer-based architecture for generalised visual object counting, termed as Counting Transformer (CounTR), which explicitly capture the similarity between image patches or with given "exemplars" with the attention mechanism;(2) We adopt a two-stage training regime, that first pre-trains the model with self-supervised learning, and followed by supervised fine-tuning;(3) We propose a simple, scalable pipeline for synthesizing training images with a large number of instances or that from different semantic categories, explicitly forcing the model to make use of the given "exemplars";(4) We conduct thorough ablation studies on the large-scale counting benchmark, e.g. FSC-147, and demonstrate state-of-the-art performance on both zero and few-shot settings.

Tengda Han, Weidi Xie, Andrew Zisserman

The objective of this paper is a temporal alignment network that ingests long term video sequences, and associated text sentences, in order to: (1) determine if a sentence is alignable with the video; and (2) if it is alignable, then determine its alignment. The challenge is to train such networks from large-scale datasets, such as HowTo100M, where the associated text sentences have significant noise, and are only weakly aligned when relevant. Apart from proposing the alignment network, we also make four contributions: (i) we describe a novel co-training method that enables to denoise and train on raw instructional videos without using manual annotation, despite the considerable noise; (ii) to benchmark the alignment performance, we manually curate a 10-hour subset of HowTo100M, totalling 80 videos, with sparse temporal descriptions. Our proposed model, trained on HowTo100M, outperforms strong baselines (CLIP, MIL-NCE) on this alignment dataset by a significant margin; (iii) we apply the trained model in the zero-shot settings to multiple downstream video understanding tasks and achieve state-of-the-art results, including text-video retrieval on YouCook2, and weakly supervised video action segmentation on Breakfast-Action; (iv) we use the automatically aligned HowTo100M annotations for end-to-end finetuning of the backbone model, and obtain improved performance on downstream action recognition tasks.

Jaesung Huh, Max Bain, Andrew Zisserman

This report presents the technical details of our submission on the EGO4D Audio-Visual (AV) Automatic Speech Recognition Challenge 2023 from the OxfordVGG team. We present WhisperX, a system for efficient speech transcription of long-form audio with word-level time alignment, along with two text normalisers which are publicly available. Our final submission obtained 56.0% of the Word Error Rate (WER) on the challenge test set, ranked 1st on the leaderboard. All baseline codes and models are available on https://github.com/m-bain/whisperX.

Liliane Momeni, Mathilde Caron, Arsha Nagrani et al.

Understanding verbs is crucial to modelling how people and objects interact with each other and the environment through space and time. Recently, state-of-the-art video-language models based on CLIP have been shown to have limited verb understanding and to rely extensively on nouns, restricting their performance in real-world video applications that require action and temporal understanding. In this work, we improve verb understanding for CLIP-based video-language models by proposing a new Verb-Focused Contrastive (VFC) framework. This consists of two main components: (1) leveraging pretrained large language models (LLMs) to create hard negatives for cross-modal contrastive learning, together with a calibration strategy to balance the occurrence of concepts in positive and negative pairs; and (2) enforcing a fine-grained, verb phrase alignment loss. Our method achieves state-of-the-art results for zero-shot performance on three downstream tasks that focus on verb understanding: video-text matching, video question-answering and video classification. To the best of our knowledge, this is the first work which proposes a method to alleviate the verb understanding problem, and does not simply highlight it.

Ragav Sachdeva, Gyungin Shin, Andrew Zisserman · oxford

Enabling engagement of manga by visually impaired individuals presents a significant challenge due to its inherently visual nature. With the goal of fostering accessibility, this paper aims to generate a dialogue transcript of a complete manga chapter, entirely automatically, with a particular emphasis on ensuring narrative consistency. This entails identifying (i) what is being said, i.e., detecting the texts on each page and classifying them into essential vs non-essential, and (ii) who is saying it, i.e., attributing each dialogue to its speaker, while ensuring the same characters are named consistently throughout the chapter. To this end, we introduce: (i) Magiv2, a model that is capable of generating high-quality chapter-wide manga transcripts with named characters and significantly higher precision in speaker diarisation over prior works; (ii) an extension of the PopManga evaluation dataset, which now includes annotations for speech-bubble tail boxes, associations of text to corresponding tails, classifications of text as essential or non-essential, and the identity for each character box; and (iii) a new character bank dataset, which comprises over 11K characters from 76 manga series, featuring 11.5K exemplar character images in total, as well as a list of chapters in which they appear. The code, trained model, and both datasets can be found at: https://github.com/ragavsachdeva/magi

Max Bain, Arsha Nagrani, Gül Varol et al.

Our goal in this paper is the adaptation of image-text models for long video retrieval. Recent works have demonstrated state-of-the-art performance in video retrieval by adopting CLIP, effectively hitchhiking on the image-text representation for video tasks. However, there has been limited success in learning temporal aggregation that outperform mean-pooling the image-level representations extracted per frame by CLIP. We find that the simple yet effective baseline of weighted-mean of frame embeddings via query-scoring is a significant improvement above all prior temporal modelling attempts and mean-pooling. In doing so, we provide an improved baseline for others to compare to and demonstrate state-of-the-art performance of this simple baseline on a suite of long video retrieval benchmarks.

Yash Bhalgat, Iro Laina, João F. Henriques et al.

Instance segmentation in 3D is a challenging task due to the lack of large-scale annotated datasets. In this paper, we show that this task can be addressed effectively by leveraging instead 2D pre-trained models for instance segmentation. We propose a novel approach to lift 2D segments to 3D and fuse them by means of a neural field representation, which encourages multi-view consistency across frames. The core of our approach is a slow-fast clustering objective function, which is scalable and well-suited for scenes with a large number of objects. Unlike previous approaches, our method does not require an upper bound on the number of objects or object tracking across frames. To demonstrate the scalability of the slow-fast clustering, we create a new semi-realistic dataset called the Messy Rooms dataset, which features scenes with up to 500 objects per scene. Our approach outperforms the state-of-the-art on challenging scenes from the ScanNet, Hypersim, and Replica datasets, as well as on our newly created Messy Rooms dataset, demonstrating the effectiveness and scalability of our slow-fast clustering method.

Junyu Xie, Weidi Xie, Andrew Zisserman

The objective of this paper is a model that is able to discover, track and segment multiple moving objects in a video. We make four contributions: First, we introduce an object-centric segmentation model with a depth-ordered layer representation. This is implemented using a variant of the transformer architecture that ingests optical flow, where each query vector specifies an object and its layer for the entire video. The model can effectively discover multiple moving objects and handle mutual occlusions; Second, we introduce a scalable pipeline for generating multi-object synthetic training data via layer compositions, that is used to train the proposed model, significantly reducing the requirements for labour-intensive annotations, and supporting Sim2Real generalisation; Third, we conduct thorough ablation studies, showing that the model is able to learn object permanence and temporal shape consistency, and is able to predict amodal segmentation masks; Fourth, we evaluate our model, trained only on synthetic data, on standard video segmentation benchmarks, DAVIS, MoCA, SegTrack, FBMS-59, and achieve state-of-the-art performance among existing methods that do not rely on any manual annotations. With test-time adaptation, we observe further performance boosts.

Prannay Kaul, Weidi Xie, Andrew Zisserman

The goal of this paper is open-vocabulary object detection (OVOD) $\unicode{x2013}$ building a model that can detect objects beyond the set of categories seen at training, thus enabling the user to specify categories of interest at inference without the need for model retraining. We adopt a standard two-stage object detector architecture, and explore three ways for specifying novel categories: via language descriptions, via image exemplars, or via a combination of the two. We make three contributions: first, we prompt a large language model (LLM) to generate informative language descriptions for object classes, and construct powerful text-based classifiers; second, we employ a visual aggregator on image exemplars that can ingest any number of images as input, forming vision-based classifiers; and third, we provide a simple method to fuse information from language descriptions and image exemplars, yielding a multi-modal classifier. When evaluating on the challenging LVIS open-vocabulary benchmark we demonstrate that: (i) our text-based classifiers outperform all previous OVOD works; (ii) our vision-based classifiers perform as well as text-based classifiers in prior work; (iii) using multi-modal classifiers perform better than either modality alone; and finally, (iv) our text-based and multi-modal classifiers yield better performance than a fully-supervised detector.

Sagar Vaze, Andrea Vedaldi, Andrew Zisserman

In this paper we tackle the problem of Generalized Category Discovery (GCD). Specifically, given a dataset with labelled and unlabelled images, the task is to cluster all images in the unlabelled subset, whether or not they belong to the labelled categories. Our first contribution is to recognize that most existing GCD benchmarks only contain labels for a single clustering of the data, making it difficult to ascertain whether models are using the available labels to solve the GCD task, or simply solving an unsupervised clustering problem. As such, we present a synthetic dataset, named 'Clevr-4', for category discovery. Clevr-4 contains four equally valid partitions of the data, i.e based on object shape, texture, color or count. To solve the task, models are required to extrapolate the taxonomy specified by the labelled set, rather than simply latching onto a single natural grouping of the data. We use this dataset to demonstrate the limitations of unsupervised clustering in the GCD setting, showing that even very strong unsupervised models fail on Clevr-4. We further use Clevr-4 to examine the weaknesses of existing GCD algorithms, and propose a new method which addresses these shortcomings, leveraging consistent findings from the representation learning literature to do so. Our simple solution, which is based on 'mean teachers' and termed $μ$GCD, substantially outperforms implemented baselines on Clevr-4. Finally, when we transfer these findings to real data on the challenging Semantic Shift Benchmark (SSB), we find that $μ$GCD outperforms all prior work, setting a new state-of-the-art. For the project webpage, see https://www.robots.ox.ac.uk/~vgg/data/clevr4/

Jaesung Huh, Jacob Chalk, Evangelos Kazakos et al.

We introduce EPIC-SOUNDS, a large-scale dataset of audio annotations capturing temporal extents and class labels within the audio stream of the egocentric videos. We propose an annotation pipeline where annotators temporally label distinguishable audio segments and describe the action that could have caused this sound. We identify actions that can be discriminated purely from audio, through grouping these free-form descriptions of audio into classes. For actions that involve objects colliding, we collect human annotations of the materials of these objects (e.g. a glass object being placed on a wooden surface), which we verify from video, discarding ambiguities. Overall, EPIC-SOUNDS includes 78.4k categorised segments of audible events and actions, distributed across 44 classes as well as 39.2k non-categorised segments. We train and evaluate state-of-the-art audio recognition and detection models on our dataset, for both audio-only and audio-visual methods. We also conduct analysis on: the temporal overlap between audio events, the temporal and label correlations between audio and visual modalities, the ambiguities in annotating materials from audio-only input, the importance of audio-only labels and the limitations of current models to understand actions that sound.

Tengda Han, Max Bain, Arsha Nagrani et al.

The objective of this paper is an automatic Audio Description (AD) model that ingests movies and outputs AD in text form. Generating high-quality movie AD is challenging due to the dependency of the descriptions on context, and the limited amount of training data available. In this work, we leverage the power of pretrained foundation models, such as GPT and CLIP, and only train a mapping network that bridges the two models for visually-conditioned text generation. In order to obtain high-quality AD, we make the following four contributions: (i) we incorporate context from the movie clip, AD from previous clips, as well as the subtitles; (ii) we address the lack of training data by pretraining on large-scale datasets, where visual or contextual information is unavailable, e.g. text-only AD without movies or visual captioning datasets without context; (iii) we improve on the currently available AD datasets, by removing label noise in the MAD dataset, and adding character naming information; and (iv) we obtain strong results on the movie AD task compared with previous methods.

Niki Amini-Naieni, Kiana Amini-Naieni, Tengda Han et al.

Our objective is open-world object counting in images, where the target object class is specified by a text description. To this end, we propose CounTX, a class-agnostic, single-stage model using a transformer decoder counting head on top of pre-trained joint text-image representations. CounTX is able to count the number of instances of any class given only an image and a text description of the target object class, and can be trained end-to-end. In addition to this model, we make the following contributions: (i) we compare the performance of CounTX to prior work on open-world object counting, and show that our approach exceeds the state of the art on all measures on the FSC-147 benchmark for methods that use text to specify the task; (ii) we present and release FSC-147-D, an enhanced version of FSC-147 with text descriptions, so that object classes can be described with more detailed language than their simple class names. FSC-147-D and the code are available at https://www.robots.ox.ac.uk/~vgg/research/countx.

Vladimir Iashin, Weidi Xie, Esa Rahtu et al.

The objective of this paper is audio-visual synchronisation of general videos 'in the wild'. For such videos, the events that may be harnessed for synchronisation cues may be spatially small and may occur only infrequently during a many seconds-long video clip, i.e. the synchronisation signal is 'sparse in space and time'. This contrasts with the case of synchronising videos of talking heads, where audio-visual correspondence is dense in both time and space. We make four contributions: (i) in order to handle longer temporal sequences required for sparse synchronisation signals, we design a multi-modal transformer model that employs 'selectors' to distil the long audio and visual streams into small sequences that are then used to predict the temporal offset between streams. (ii) We identify artefacts that can arise from the compression codecs used for audio and video and can be used by audio-visual models in training to artificially solve the synchronisation task. (iii) We curate a dataset with only sparse in time and space synchronisation signals; and (iv) the effectiveness of the proposed model is shown on both dense and sparse datasets quantitatively and qualitatively. Project page: v-iashin.github.io/SparseSync

Chuhan Zhang, Ankush Gupta, Andrew Zisserman

The objective of this work is to learn an object-centric video representation, with the aim of improving transferability to novel tasks, i.e., tasks different from the pre-training task of action classification. To this end, we introduce a new object-centric video recognition model based on a transformer architecture. The model learns a set of object-centric summary vectors for the video, and uses these vectors to fuse the visual and spatio-temporal trajectory 'modalities' of the video clip. We also introduce a novel trajectory contrast loss to further enhance objectness in these summary vectors. With experiments on four datasets -- SomethingSomething-V2, SomethingElse, Action Genome and EpicKitchens -- we show that the object-centric model outperforms prior video representations (both object-agnostic and object-aware), when: (1) classifying actions on unseen objects and unseen environments; (2) low-shot learning of novel classes; (3) linear probe to other downstream tasks; as well as (4) for standard action classification.

Jaesung Huh, Andrew Brown, Jee-weon Jung et al.

This paper summarises the findings from the VoxCeleb Speaker Recognition Challenge 2022 (VoxSRC-22), which was held in conjunction with INTERSPEECH 2022. The goal of this challenge was to evaluate how well state-of-the-art speaker recognition systems can diarise and recognise speakers from speech obtained "in the wild". The challenge consisted of: (i) the provision of publicly available speaker recognition and diarisation data from YouTube videos together with ground truth annotation and standardised evaluation software; and (ii) a public challenge and hybrid workshop held at INTERSPEECH 2022. We describe the four tracks of our challenge along with the baselines, methods, and results. We conclude with a discussion on the new domain-transfer focus of VoxSRC-22, and on the progression of the challenge from the previous three editions.

Guanqi Zhan, Weidi Xie, Andrew Zisserman

Detecting occluded objects still remains a challenge for state-of-the-art object detectors. The objective of this work is to improve the detection for such objects, and thereby improve the overall performance of a modern object detector. To this end we make the following four contributions: (1) We propose a simple 'plugin' module for the detection head of two-stage object detectors to improve the recall of partially occluded objects. The module predicts a tri-layer of segmentation masks for the target object, the occluder and the occludee, and by doing so is able to better predict the mask of the target object. (2) We propose a scalable pipeline for generating training data for the module by using amodal completion of existing object detection and instance segmentation training datasets to establish occlusion relationships. (3) We also establish a COCO evaluation dataset to measure the recall performance of partially occluded and separated objects. (4) We show that the plugin module inserted into a two-stage detector can boost the performance significantly, by only fine-tuning the detection head, and with additional improvements if the entire architecture is fine-tuned. COCO results are reported for Mask R-CNN with Swin-T or Swin-S backbones, and Cascade Mask R-CNN with a Swin-B backbone.

Relja Arandjelović, Alex Andonian, Arthur Mensch et al.

The core problem in zero-shot open vocabulary detection is how to align visual and text features, so that the detector performs well on unseen classes. Previous approaches train the feature pyramid and detection head from scratch, which breaks the vision-text feature alignment established during pretraining, and struggles to prevent the language model from forgetting unseen classes. We propose three methods to alleviate these issues. Firstly, a simple scheme is used to augment the text embeddings which prevents overfitting to a small number of classes seen during training, while simultaneously saving memory and computation. Secondly, the feature pyramid network and the detection head are modified to include trainable gated shortcuts, which encourages vision-text feature alignment and guarantees it at the start of detection training. Finally, a self-training approach is used to leverage a larger corpus of image-text pairs thus improving detection performance on classes with no human annotated bounding boxes. Our three methods are evaluated on the zero-shot version of the LVIS benchmark, each of them showing clear and significant benefits. Our final network achieves the new stateof-the-art on the mAP-all metric and demonstrates competitive performance for mAP-rare, as well as superior transfer to COCO and Objects365.

Hala Lamdouar, Weidi Xie, Andrew Zisserman

Not all camouflages are equally effective, as even a partially visible contour or a slight color difference can make the animal stand out and break its camouflage. In this paper, we address the question of what makes a camouflage successful, by proposing three scores for automatically assessing its effectiveness. In particular, we show that camouflage can be measured by the similarity between background and foreground features and boundary visibility. We use these camouflage scores to assess and compare all available camouflage datasets. We also incorporate the proposed camouflage score into a generative model as an auxiliary loss and show that effective camouflage images or videos can be synthesised in a scalable manner. The generated synthetic dataset is used to train a transformer-based model for segmenting camouflaged animals in videos. Experimentally, we demonstrate state-of-the-art camouflage breaking performance on the public MoCA-Mask benchmark.

Rhydian Windsor, Amir Jamaludin, Timor Kadir et al.

This paper explores training medical vision-language models (VLMs) -- where the visual and language inputs are embedded into a common space -- with a particular focus on scenarios where training data is limited, as is often the case in clinical datasets. We explore several candidate methods to improve low-data performance, including: (i) adapting generic pre-trained models to novel image and text domains (i.e. medical imaging and reports) via unimodal self-supervision; (ii) using local (e.g. GLoRIA) & global (e.g. InfoNCE) contrastive loss functions as well as a combination of the two; (iii) extra supervision during VLM training, via: (a) image- and text-only self-supervision, and (b) creating additional positive image-text pairs for training through augmentation and nearest-neighbour search. Using text-to-image retrieval as a benchmark, we evaluate the performance of these methods with variable sized training datasets of paired chest X-rays and radiological reports. Combined, they significantly improve retrieval compared to fine-tuning CLIP, roughly equivalent to training with the data. A similar pattern is found in the downstream task classification of CXR-related conditions with our method outperforming CLIP and also BioVIL, a strong CXR VLM benchmark, in the zero-shot and linear probing settings. We conclude with a set of recommendations for researchers aiming to train vision-language models on other medical imaging modalities when training data is scarce. To facilitate further research, we will make our code and models publicly available.

Tengda Han, Weidi Xie, Andrew Zisserman

The objective of this paper is an efficient training method for video tasks. We make three contributions: (1) We propose Turbo training, a simple and versatile training paradigm for Transformers on multiple video tasks. (2) We illustrate the advantages of Turbo training on action classification, video-language representation learning, and long-video activity classification, showing that Turbo training can largely maintain competitive performance while achieving almost 4X speed-up and significantly less memory consumption. (3) Turbo training enables long-schedule video-language training and end-to-end long-video training, delivering competitive or superior performance than previous works, which were infeasible to train under limited resources.

Rhydian Windsor, Amir Jamaludin, Timor Kadir et al.

This paper proposes a novel transformer-based model architecture for medical imaging problems involving analysis of vertebrae. It considers two applications of such models in MR images: (a) detection of spinal metastases and the related conditions of vertebral fractures and metastatic cord compression, (b) radiological grading of common degenerative changes in intervertebral discs. Our contributions are as follows: (i) We propose a Spinal Context Transformer (SCT), a deep-learning architecture suited for the analysis of repeated anatomical structures in medical imaging such as vertebral bodies (VBs). Unlike previous related methods, SCT considers all VBs as viewed in all available image modalities together, making predictions for each based on context from the rest of the spinal column and all available imaging modalities. (ii) We apply the architecture to a novel and important task: detecting spinal metastases and the related conditions of cord compression and vertebral fractures/collapse from multi-series spinal MR scans. This is done using annotations extracted from free-text radiological reports as opposed to bespoke annotation. However, the resulting model shows strong agreement with vertebral-level bespoke radiologist annotations on the test set. (iii) We also apply SCT to an existing problem: radiological grading of inter-vertebral discs (IVDs) in lumbar MR scans for common degenerative changes.We show that by considering the context of vertebral bodies in the image, SCT improves the accuracy for several gradings compared to previously published model.

Olivia Wiles, Joao Carreira, Iain Barr et al.

Experience and reasoning occur across multiple temporal scales: milliseconds, seconds, hours or days. The vast majority of computer vision research, however, still focuses on individual images or short videos lasting only a few seconds. This is because handling longer videos require more scalable approaches even to process them. In this work, we propose a framework enabling research on hour-long videos with the same hardware that can now process second-long videos. We replace standard video compression, e.g. JPEG, with neural compression and show that we can directly feed compressed videos as inputs to regular video networks. Operating on compressed videos improves efficiency at all pipeline levels -- data transfer, speed and memory -- making it possible to train models faster and on much longer videos. Processing compressed signals has, however, the downside of precluding standard augmentation techniques if done naively. We address that by introducing a small network that can apply transformations to latent codes corresponding to commonly used augmentations in the original video space. We demonstrate that with our compressed vision pipeline, we can train video models more efficiently on popular benchmarks such as Kinetics600 and COIN. We also perform proof-of-concept experiments with new tasks defined over hour-long videos at standard frame rates. Processing such long videos is impossible without using compressed representation.

Rhydian Windsor, Amir Jamaludin, Timor Kadir et al.

This technical report presents SpineNetV2, an automated tool which: (i) detects and labels vertebral bodies in clinical spinal magnetic resonance (MR) scans across a range of commonly used sequences; and (ii) performs radiological grading of lumbar intervertebral discs in T2-weighted scans for a range of common degenerative changes. SpineNetV2 improves over the original SpineNet software in two ways: (1) The vertebral body detection stage is significantly faster, more accurate and works across a range of fields-of-view (as opposed to just lumbar scans). (2) Radiological grading adopts a more powerful architecture, adding several new grading schemes without loss in performance. A demo of the software is available at the project website: http://zeus.robots.ox.ac.uk/spinenet2/.

Tengda Han, Max Bain, Arsha Nagrani et al.

Audio Description (AD) is the task of generating descriptions of visual content, at suitable time intervals, for the benefit of visually impaired audiences. For movies, this presents notable challenges -- AD must occur only during existing pauses in dialogue, should refer to characters by name, and ought to aid understanding of the storyline as a whole. To this end, we develop a new model for automatically generating movie AD, given CLIP visual features of the frames, the cast list, and the temporal locations of the speech; addressing all three of the 'who', 'when', and 'what' questions: (i) who -- we introduce a character bank consisting of the character's name, the actor that played the part, and a CLIP feature of their face, for the principal cast of each movie, and demonstrate how this can be used to improve naming in the generated AD; (ii) when -- we investigate several models for determining whether an AD should be generated for a time interval or not, based on the visual content of the interval and its neighbours; and (iii) what -- we implement a new vision-language model for this task, that can ingest the proposals from the character bank, whilst conditioning on the visual features using cross-attention, and demonstrate how this improves over previous architectures for AD text generation in an apples-to-apples comparison.

Bruno Korbar, Andrew Zisserman

Multiple-object tracking (MOT) is a challenging task that requires simultaneous reasoning about location, appearance, and identity of the objects in the scene over time. Our aim in this paper is to move beyond tracking-by-detection approaches, that perform well on datasets where the object classes are known, to class-agnostic tracking that performs well also for unknown object classes.To this end, we make the following three contributions: first, we introduce {\em semantic detector queries} that enable an object to be localized by specifying its approximate position, or its appearance, or both; second, we use these queries within an auto-regressive framework for tracking, and propose a multi-query tracking transformer (\textit{MQT}) model for simultaneous tracking and appearance-based re-identification (reID) based on the transformer architecture with deformable attention. This formulation allows the tracker to operate in a class-agnostic manner, and the model can be trained end-to-end; finally, we demonstrate that \textit{MQT} performs competitively on standard MOT benchmarks, outperforms all baselines on generalised-MOT, and generalises well to a much harder tracking problems such as tracking any object on the TAO dataset.

Yash Bhalgat, Joao F. Henriques, Andrew Zisserman

Transformers are powerful visual learners, in large part due to their conspicuous lack of manually-specified priors. This flexibility can be problematic in tasks that involve multiple-view geometry, due to the near-infinite possible variations in 3D shapes and viewpoints (requiring flexibility), and the precise nature of projective geometry (obeying rigid laws). To resolve this conundrum, we propose a "light touch" approach, guiding visual Transformers to learn multiple-view geometry but allowing them to break free when needed. We achieve this by using epipolar lines to guide the Transformer's cross-attention maps, penalizing attention values outside the epipolar lines and encouraging higher attention along these lines since they contain geometrically plausible matches. Unlike previous methods, our proposal does not require any camera pose information at test-time. We focus on pose-invariant object instance retrieval, where standard Transformer networks struggle, due to the large differences in viewpoint between query and retrieved images. Experimentally, our method outperforms state-of-the-art approaches at object retrieval, without needing pose information at test-time.

Guanqi Zhan, Chuanxia Zheng, Weidi Xie et al.

Our objective in this paper is to probe large vision models to determine to what extent they 'understand' different physical properties of the 3D scene depicted in an image. To this end, we make the following contributions: (i) We introduce a general and lightweight protocol to evaluate whether features of an off-the-shelf large vision model encode a number of physical 'properties' of the 3D scene, by training discriminative classifiers on the features for these properties. The probes are applied on datasets of real images with annotations for the property. (ii) We apply this protocol to properties covering scene geometry, scene material, support relations, lighting, and view-dependent measures, and large vision models including CLIP, DINOv1, DINOv2, VQGAN, Stable Diffusion. (iii) We find that features from Stable Diffusion and DINOv2 are good for discriminative learning of a number of properties, including scene geometry, support relations, shadows and depth, but less performant for occlusion and material, while outperforming DINOv1, CLIP and VQGAN for all properties. (iv) It is observed that different time steps of Stable Diffusion features, as well as different transformer layers of DINO/CLIP/VQGAN, are good at different properties, unlocking potential applications of 3D physical understanding.

Yash Bhalgat, Vadim Tschernezki, Iro Laina et al.

Egocentric videos present unique challenges for 3D scene understanding due to rapid camera motion, frequent object occlusions, and limited object visibility. This paper introduces a novel approach to instance segmentation and tracking in first-person video that leverages 3D awareness to overcome these obstacles. Our method integrates scene geometry, 3D object centroid tracking, and instance segmentation to create a robust framework for analyzing dynamic egocentric scenes. By incorporating spatial and temporal cues, we achieve superior performance compared to state-of-the-art 2D approaches. Extensive evaluations on the challenging EPIC Fields dataset demonstrate significant improvements across a range of tracking and segmentation consistency metrics. Specifically, our method outperforms the next best performing approach by $7$ points in Association Accuracy (AssA) and $4.5$ points in IDF1 score, while reducing the number of ID switches by $73\%$ to $80\%$ across various object categories. Leveraging our tracked instance segmentations, we showcase downstream applications in 3D object reconstruction and amodal video object segmentation in these egocentric settings.

Amir Jamaludin, Timor Kadir, Emma Clark et al.

Our objective in this paper is to estimate spine curvature in DXA scans. To this end we first train a neural network to predict the middle spine curve in the scan, and then use an integral-based method to determine the curvature along the spine curve. We use the curvature to compare to the standard angle scoliosis measure obtained using the DXA Scoliosis Method (DSM). The performance improves over the prior work of Jamaludin et al. 2018. We show that the maximum curvature can be used as a scoring function for ordering the severity of spinal deformation.

Ragav Sachdeva, Andrew Zisserman

We live in a dynamic world where things change all the time. Given two images of the same scene, being able to automatically detect the changes in them has practical applications in a variety of domains. In this paper, we tackle the change detection problem with the goal of detecting "object-level" changes in an image pair despite differences in their viewpoint and illumination. To this end, we make the following four contributions: (i) we propose a scalable methodology for obtaining a large-scale change detection training dataset by leveraging existing object segmentation benchmarks; (ii) we introduce a co-attention based novel architecture that is able to implicitly determine correspondences between an image pair and find changes in the form of bounding box predictions; (iii) we contribute four evaluation datasets that cover a variety of domains and transformations, including synthetic image changes, real surveillance images of a 3D scene, and synthetic 3D scenes with camera motion; (iv) we evaluate our model on these four datasets and demonstrate zero-shot and beyond training transformation generalization.

Junyu Xie, Tengda Han, Max Bain et al.

Our objective is to generate Audio Descriptions (ADs) for both movies and TV series in a training-free manner. We use the power of off-the-shelf Visual-Language Models (VLMs) and Large Language Models (LLMs), and develop visual and text prompting strategies for this task. Our contributions are three-fold: (i) We demonstrate that a VLM can successfully name and refer to characters if directly prompted with character information through visual indications without requiring any fine-tuning; (ii) A two-stage process is developed to generate ADs, with the first stage asking the VLM to comprehensively describe the video, followed by a second stage utilising a LLM to summarise dense textual information into one succinct AD sentence; (iii) A new dataset for TV audio description is formulated. Our approach, named AutoAD-Zero, demonstrates outstanding performance (even competitive with some models fine-tuned on ground truth ADs) in AD generation for both movies and TV series, achieving state-of-the-art CRITIC scores.

Skanda Koppula, Ignacio Rocco, Yi Yang et al.

We introduce a new benchmark, TAPVid-3D, for evaluating the task of long-range Tracking Any Point in 3D (TAP-3D). While point tracking in two dimensions (TAP) has many benchmarks measuring performance on real-world videos, such as TAPVid-DAVIS, three-dimensional point tracking has none. To this end, leveraging existing footage, we build a new benchmark for 3D point tracking featuring 4,000+ real-world videos, composed of three different data sources spanning a variety of object types, motion patterns, and indoor and outdoor environments. To measure performance on the TAP-3D task, we formulate a collection of metrics that extend the Jaccard-based metric used in TAP to handle the complexities of ambiguous depth scales across models, occlusions, and multi-track spatio-temporal smoothness. We manually verify a large sample of trajectories to ensure correct video annotations, and assess the current state of the TAP-3D task by constructing competitive baselines using existing tracking models. We anticipate this benchmark will serve as a guidepost to improve our ability to understand precise 3D motion and surface deformation from monocular video. Code for dataset download, generation, and model evaluation is available at https://tapvid3d.github.io

Prasanna Sridhar, Horace Lee, David M. S. Pinto et al.

In this paper, we present WISE, an open-source audiovisual search engine which integrates a range of multimodal retrieval capabilities into a single, practical tool accessible to users without machine learning expertise. WISE supports natural-language and reverse-image queries at both the scene level (e.g. empty street) and object level (e.g. horse) across images and videos; face-based search for specific individuals; audio retrieval of acoustic events using text (e.g. wood creak) or an audio file; search over automatically transcribed speech; and filtering by user-provided metadata. Rich insights can be obtained by combining queries across modalities -- for example, retrieving German trains from a historical archive by applying the object query "train" and the metadata query "Germany", or searching for a face in a place. By employing vector search techniques, WISE can scale to support efficient retrieval over millions of images or thousands of hours of video. Its modular architecture facilitates the integration of new models. WISE can be deployed locally for private or sensitive collections, and has been applied to various real-world use cases. Our code is open-source and available at https://gitlab.com/vgg/wise/wise.

Niki Amini-Naieni, Andrew Zisserman

The flexibility and accuracy of methods for automatically counting objects in images and videos are limited by the way the object can be specified. While existing methods allow users to describe the target object with text and visual examples, the visual examples must be manually annotated inside the image, and there is no way to specify what not to count. To address these gaps, we introduce novel capabilities that expand how the target object can be specified. Specifically, we extend the prompt to enable what not to count to be described with text and/or visual examples, introduce the concept of `pseudo-exemplars' that automate the annotation of visual examples at inference, and extend counting models to accept visual examples from both natural and synthetic external images. We also use our new counting model, CountGD++, as a vision expert agent for an LLM. Together, these contributions expand the prompt flexibility of multi-modal open-world counting and lead to significant improvements in accuracy, efficiency, and generalization across multiple datasets. Code is available at https://github.com/niki-amini-naieni/CountGDPlusPlus.

Sindhu B Hegde, K R Prajwal, Andrew Zisserman

While humans naturally gesture during speech, only a sparse subset of these movements are visually depictive and semantically linked to specific spoken words. Current multimodal models struggle to capture these semantic co-speech gestures, heavily bottlenecked by a lack of precisely annotated training data. To address this, we introduce the Gesture Recognition in the Wild (GRW) dataset, the first large-scale benchmark designed to map unconstrained human gestures to specific words with frame-accurate temporal boundaries. Comprising 156,688 manually annotated video clips, GRW spans a highly diverse 150-word taxonomy of physical actions, spatial descriptors, and abstract concepts. We leverage GRW to train video models to (a) classify gestures as semantic or not, (b) recognize the word corresponding to a co-speech gesture, and (c) temporally localize the gesture. We also use GRW to establish benchmarks for these three tasks.

Debidatta Dwibedi, Yusuf Aytar, Jonathan Tompson et al.

We introduce a dataset of annotations of temporal repetitions in videos. The dataset, OVR (pronounced as over), contains annotations for over 72K videos, with each annotation specifying the number of repetitions, the start and end time of the repetitions, and also a free-form description of what is repeating. The annotations are provided for videos sourced from Kinetics and Ego4D, and consequently cover both Exo and Ego viewing conditions, with a huge variety of actions and activities. Moreover, OVR is almost an order of magnitude larger than previous datasets for video repetition. We also propose a baseline transformer-based counting model, OVRCounter, that can localise and count repetitions in videos that are up to 320 frames long. The model is trained and evaluated on the OVR dataset, and its performance assessed with and without using text to specify the target class to count. The performance is also compared to a prior repetition counting model. The dataset is available for download at: https://sites.google.com/view/openvocabreps/

Junyu Xie, Tengda Han, Weidi Xie et al.

Moving Object Segmentation (MOS) aims to discover, segment, and track objects that move independently of the camera. Current MOS methods, however, exhibit two fundamental limitations: they rely on pre-computed 2D auxiliary modalities such as optical flow or point trajectories that lack 3D geometric information, and they treat motion as a sequence-level attribute, overlooking the instantaneous motion state of each object. We address both by grounding MOS in 3D space and time, and propose GMOS, a framework that operates directly on RGB video to produce 3D-aware, temporally fine-grained segmentation of multiple moving objects, alongside a foreground--background variant GMOS-S for faster deployment. To support training and evaluation in this regime, we curate GMOS-2K, a dataset of 2,210 real-world videos with per-object temporal motion annotations drawn from five established Video Object Segmentation (VOS) benchmarks, and formalise MOS-I ("I" for instantaneous), a temporally fine-grained evaluation protocol with three complementary metrics. GMOS achieves state-of-the-art results across MOS, MOS-I, and Unsupervised VOS benchmarks, while running significantly faster than prior multi-object MOS methods and supporting online inference for streaming deployment.

Niki Amini-Naieni, Tengda Han, Andrew Zisserman

The goal of this paper is to improve the generality and accuracy of open-vocabulary object counting in images. To improve the generality, we repurpose an open-vocabulary detection foundation model (GroundingDINO) for the counting task, and also extend its capabilities by introducing modules to enable specifying the target object to count by visual exemplars. In turn, these new capabilities - being able to specify the target object by multi-modalites (text and exemplars) - lead to an improvement in counting accuracy. We make three contributions: First, we introduce the first open-world counting model, CountGD, where the prompt can be specified by a text description or visual exemplars or both; Second, we show that the performance of the model significantly improves the state of the art on multiple counting benchmarks - when using text only, CountGD is comparable to or outperforms all previous text-only works, and when using both text and visual exemplars, we outperform all previous models; Third, we carry out a preliminary study into different interactions between the text and visual exemplar prompts, including the cases where they reinforce each other and where one restricts the other. The code and an app to test the model are available at https://www.robots.ox.ac.uk/~vgg/research/countgd/.

Chuhan Zhang, Ankush Gupta, Andrew Zisserman

We introduce an object-aware decoder for improving the performance of spatio-temporal representations on ego-centric videos. The key idea is to enhance object-awareness during training by tasking the model to predict hand positions, object positions, and the semantic label of the objects using paired captions when available. At inference time the model only requires RGB frames as inputs, and is able to track and ground objects (although it has not been trained explicitly for this). We demonstrate the performance of the object-aware representations learnt by our model, by: (i) evaluating it for strong transfer, i.e. through zero-shot testing, on a number of downstream video-text retrieval and classification benchmarks; and (ii) by using the representations learned as input for long-term video understanding tasks (e.g. Episodic Memory in Ego4D). In all cases the performance improves over the state of the art -- even compared to networks trained with far larger batch sizes. We also show that by using noisy image-level detection as pseudo-labels in training, the model learns to provide better bounding boxes using video consistency, as well as grounding the words in the associated text descriptions. Overall, we show that the model can act as a drop-in replacement for an ego-centric video model to improve performance through visual-text grounding.

Jaesung Huh, Joon Son Chung, Arsha Nagrani et al.

The VoxCeleb Speaker Recognition Challenges (VoxSRC) were a series of challenges and workshops that ran annually from 2019 to 2023. The challenges primarily evaluated the tasks of speaker recognition and diarisation under various settings including: closed and open training data; as well as supervised, self-supervised, and semi-supervised training for domain adaptation. The challenges also provided publicly available training and evaluation datasets for each task and setting, with new test sets released each year. In this paper, we provide a review of these challenges that covers: what they explored; the methods developed by the challenge participants and how these evolved; and also the current state of the field for speaker verification and diarisation. We chart the progress in performance over the five installments of the challenge on a common evaluation dataset and provide a detailed analysis of how each year's special focus affected participants' performance. This paper is aimed both at researchers who want an overview of the speaker recognition and diarisation field, and also at challenge organisers who want to benefit from the successes and avoid the mistakes of the VoxSRC challenges. We end with a discussion of the current strengths of the field and open challenges. Project page : https://mm.kaist.ac.kr/datasets/voxceleb/voxsrc/workshop.html