Andrew Gilbert

Sarina Thomas, Andrew Gilbert, Guy Ben-Yosef

Accurate and consistent predictions of echocardiography parameters are important for cardiovascular diagnosis and treatment. In particular, segmentations of the left ventricle can be used to derive ventricular volume, ejection fraction (EF) and other relevant measurements. In this paper we propose a new automated method called EchoGraphs for predicting ejection fraction and segmenting the left ventricle by detecting anatomical keypoints. Models for direct coordinate regression based on Graph Convolutional Networks (GCNs) are used to detect the keypoints. GCNs can learn to represent the cardiac shape based on local appearance of each keypoint, as well as global spatial and temporal structures of all keypoints combined. We evaluate our EchoGraphs model on the EchoNet benchmark dataset. Compared to semantic segmentation, GCNs show accurate segmentation and improvements in robustness and inference runtime. EF is computed simultaneously to segmentations and our method also obtains state-of-the-art ejection fraction estimation. Source code is available online: https://github.com/guybenyosef/EchoGraphs.

Dan Ruta, Gemma Canet Tarrés, Andrew Gilbert et al.

Neural Style Transfer (NST) is the field of study applying neural techniques to modify the artistic appearance of a content image to match the style of a reference style image. Traditionally, NST methods have focused on texture-based image edits, affecting mostly low level information and keeping most image structures the same. However, style-based deformation of the content is desirable for some styles, especially in cases where the style is abstract or the primary concept of the style is in its deformed rendition of some content. With the recent introduction of diffusion models, such as Stable Diffusion, we can access far more powerful image generation techniques, enabling new possibilities. In our work, we propose using this new class of models to perform style transfer while enabling deformable style transfer, an elusive capability in previous models. We show how leveraging the priors of these models can expose new artistic controls at inference time, and we document our findings in exploring this new direction for the field of style transfer.

Dan Ruta, Andrew Gilbert, Pranav Aggarwal et al.

We present StyleBabel, a unique open access dataset of natural language captions and free-form tags describing the artistic style of over 135K digital artworks, collected via a novel participatory method from experts studying at specialist art and design schools. StyleBabel was collected via an iterative method, inspired by `Grounded Theory': a qualitative approach that enables annotation while co-evolving a shared language for fine-grained artistic style attribute description. We demonstrate several downstream tasks for StyleBabel, adapting the recent ALADIN architecture for fine-grained style similarity, to train cross-modal embeddings for: 1) free-form tag generation; 2) natural language description of artistic style; 3) fine-grained text search of style. To do so, we extend ALADIN with recent advances in Visual Transformer (ViT) and cross-modal representation learning, achieving a state of the art accuracy in fine-grained style retrieval.

Dan Ruta, Andrew Gilbert, John Collomosse et al.

Style transfer is the task of reproducing the semantic contents of a source image in the artistic style of a second target image. In this paper, we present NeAT, a new state-of-the art feed-forward style transfer method. We re-formulate feed-forward style transfer as image editing, rather than image generation, resulting in a model which improves over the state-of-the-art in both preserving the source content and matching the target style. An important component of our model's success is identifying and fixing "style halos", a commonly occurring artefact across many style transfer techniques. In addition to training and testing on standard datasets, we introduce the BBST-4M dataset, a new, large scale, high resolution dataset of 4M images. As a component of curating this data, we present a novel model able to classify if an image is stylistic. We use BBST-4M to improve and measure the generalization of NeAT across a huge variety of styles. Not only does NeAT offer state-of-the-art quality and generalization, it is designed and trained for fast inference at high resolution.

Kar Balan, Shruti Agarwal, Simon Jenni et al.

We present EKILA; a decentralized framework that enables creatives to receive recognition and reward for their contributions to generative AI (GenAI). EKILA proposes a robust visual attribution technique and combines this with an emerging content provenance standard (C2PA) to address the problem of synthetic image provenance -- determining the generative model and training data responsible for an AI-generated image. Furthermore, EKILA extends the non-fungible token (NFT) ecosystem to introduce a tokenized representation for rights, enabling a triangular relationship between the asset's Ownership, Rights, and Attribution (ORA). Leveraging the ORA relationship enables creators to express agency over training consent and, through our attribution model, to receive apportioned credit, including royalty payments for the use of their assets in GenAI.

Gemma Canet Tarrés, Zhe Lin, Zhifei Zhang et al.

Compositing an object into an image involves multiple non-trivial sub-tasks such as object placement and scaling, color/lighting harmonization, viewpoint/geometry adjustment, and shadow/reflection generation. Recent generative image compositing methods leverage diffusion models to handle multiple sub-tasks at once. However, existing models face limitations due to their reliance on masking the original object during training, which constrains their generation to the input mask. Furthermore, obtaining an accurate input mask specifying the location and scale of the object in a new image can be highly challenging. To overcome such limitations, we define a novel problem of unconstrained generative object compositing, i.e., the generation is not bounded by the mask, and train a diffusion-based model on a synthesized paired dataset. Our first-of-its-kind model is able to generate object effects such as shadows and reflections that go beyond the mask, enhancing image realism. Additionally, if an empty mask is provided, our model automatically places the object in diverse natural locations and scales, accelerating the compositing workflow. Our model outperforms existing object placement and compositing models in various quality metrics and user studies.

Kar Balan, Alex Black, Simon Jenni et al.

We present DECORAIT; a decentralized registry through which content creators may assert their right to opt in or out of AI training as well as receive reward for their contributions. Generative AI (GenAI) enables images to be synthesized using AI models trained on vast amounts of data scraped from public sources. Model and content creators who may wish to share their work openly without sanctioning its use for training are thus presented with a data governance challenge. Further, establishing the provenance of GenAI training data is important to creatives to ensure fair recognition and reward for their such use. We report a prototype of DECORAIT, which explores hierarchical clustering and a combination of on/off-chain storage to create a scalable decentralized registry to trace the provenance of GenAI training data in order to determine training consent and reward creatives who contribute that data. DECORAIT combines distributed ledger technology (DLT) with visual fingerprinting, leveraging the emerging C2PA (Coalition for Content Provenance and Authenticity) standard to create a secure, open registry through which creatives may express consent and data ownership for GenAI.

Dan Ruta, Andrew Gilbert, Saeid Motiian et al.

We present HyperNST; a neural style transfer (NST) technique for the artistic stylization of images, based on Hyper-networks and the StyleGAN2 architecture. Our contribution is a novel method for inducing style transfer parameterized by a metric space, pre-trained for style-based visual search (SBVS). We show for the first time that such space may be used to drive NST, enabling the application and interpolation of styles from an SBVS system. The technical contribution is a hyper-network that predicts weight updates to a StyleGAN2 pre-trained over a diverse gamut of artistic content (portraits), tailoring the style parameterization on a per-region basis using a semantic map of the facial regions. We show HyperNST to exceed state of the art in content preservation for our stylized content while retaining good style transfer performance.

Soon Yau Cheong, Armin Mustafa, Andrew Gilbert

Text-to-image models (T2I) such as StableDiffusion have been used to generate high quality images of people. However, due to the random nature of the generation process, the person has a different appearance e.g. pose, face, and clothing, despite using the same text prompt. The appearance inconsistency makes T2I unsuitable for pose transfer. We address this by proposing a multimodal diffusion model that accepts text, pose, and visual prompting. Our model is the first unified method to perform all person image tasks - generation, pose transfer, and mask-less edit. We also pioneer using small dimensional 3D body model parameters directly to demonstrate new capability - simultaneous pose and camera view interpolation while maintaining the person's appearance.

Dan Ruta, Gemma Canet Tarres, Alexander Black et al.

Representation learning aims to discover individual salient features of a domain in a compact and descriptive form that strongly identifies the unique characteristics of a given sample respective to its domain. Existing works in visual style representation literature have tried to disentangle style from content during training explicitly. A complete separation between these has yet to be fully achieved. Our paper aims to learn a representation of visual artistic style more strongly disentangled from the semantic content depicted in an image. We use Neural Style Transfer (NST) to measure and drive the learning signal and achieve state-of-the-art representation learning on explicitly disentangled metrics. We show that strongly addressing the disentanglement of style and content leads to large gains in style-specific metrics, encoding far less semantic information and achieving state-of-the-art accuracy in downstream multimodal applications.

Edward Fish, Jon Weinbren, Andrew Gilbert

Pure vision transformer architectures are highly effective for short video classification and action recognition tasks. However, due to the quadratic complexity of self attention and lack of inductive bias, transformers are resource intensive and suffer from data inefficiencies. Long form video understanding tasks amplify data and memory efficiency problems in transformers making current approaches unfeasible to implement on data or memory restricted domains. This paper introduces an efficient Spatio-Temporal Attention Network (STAN) which uses a two-stream transformer architecture to model dependencies between static image features and temporal contextual features. Our proposed approach can classify videos up to two minutes in length on a single GPU, is data efficient, and achieves SOTA performance on several long video understanding tasks.

Soon Yau Cheong, Armin Mustafa, Andrew Gilbert

Transformers have recently been shown to generate high quality images from text input. However, the existing method of pose conditioning using skeleton image tokens is computationally inefficient and generate low quality images. Therefore we propose a new method; Keypoint Pose Encoding (KPE); KPE is 10 times more memory efficient and over 73% faster at generating high quality images from text input conditioned on the pose. The pose constraint improves the image quality and reduces errors on body extremities such as arms and legs. The additional benefits include invariance to changes in the target image domain and image resolution, making it easily scalable to higher resolution images. We demonstrate the versatility of KPE by generating photorealistic multiperson images derived from the DeepFashion dataset. We also introduce a evaluation method People Count Error (PCE) that is effective in detecting error in generated human images.

Violeta Menéndez González, Andrew Gilbert, Graeme Phillipson et al.

This paper proposes Sparse View Synthesis. This is a view synthesis problem where the number of reference views is limited, and the baseline between target and reference view is significant. Under these conditions, current radiance field methods fail catastrophically due to inescapable artifacts such 3D floating blobs, blurring and structural duplication, whenever the number of reference views is limited, or the target view diverges significantly from the reference views. Advances in network architecture and loss regularisation are unable to satisfactorily remove these artifacts. The occlusions within the scene ensure that the true contents of these regions is simply not available to the model. In this work, we instead focus on hallucinating plausible scene contents within such regions. To this end we unify radiance field models with adversarial learning and perceptual losses. The resulting system provides up to 60% improvement in perceptual accuracy compared to current state-of-the-art radiance field models on this problem.

Mona Ahmadian, Frank Guerin, Andrew Gilbert

Self-supervised learning (SSL) techniques have recently produced outstanding results in learning visual representations from unlabeled videos. Despite the importance of motion in supervised learning techniques for action recognition, SSL methods often do not explicitly consider motion information in videos. To address this issue, we propose MOFO (MOtion FOcused), a novel SSL method for focusing representation learning on the motion area of a video, for action recognition. MOFO automatically detects motion areas in videos and uses these to guide the self-supervision task. We use a masked autoencoder which randomly masks out a high proportion of the input sequence; we force a specified percentage of the inside of the motion area to be masked and the remainder from outside. We further incorporate motion information into the finetuning step to emphasise motion in the downstream task. We demonstrate that our motion-focused innovations can significantly boost the performance of the currently leading SSL method (VideoMAE) for action recognition. Our method improves the recent self-supervised Vision Transformer (ViT), VideoMAE, by achieving +2.6%, +2.1%, +1.3% accuracy on Epic-Kitchens verb, noun and action classification, respectively, and +4.7% accuracy on Something-Something V2 action classification. Our proposed approach significantly improves the performance of the current SSL method for action recognition, indicating the importance of explicitly encoding motion in SSL.

Xu Dong, Xinran Liu, Wanqing Li et al.

Long-term Action Quality Assessment (AQA) evaluates the execution of activities in videos. However, the length presents challenges in fine-grained interpretability, with current AQA methods typically producing a single score by averaging clip features, lacking detailed semantic meanings of individual clips. Long-term videos pose additional difficulty due to the complexity and diversity of actions, exacerbating interpretability challenges. While query-based transformer networks offer promising long-term modeling capabilities, their interpretability in AQA remains unsatisfactory due to a phenomenon we term Temporal Skipping, where the model skips self-attention layers to prevent output degradation. To address this, we propose an attention loss function and a query initialization method to enhance performance and interpretability. Additionally, we introduce a weight-score regression module designed to approximate the scoring patterns observed in human judgments and replace conventional single-score regression, improving the rationality of interpretability. Our approach achieves state-of-the-art results on three real-world, long-term AQA benchmarks. Our code is available at: https://github.com/dx199771/Interpretability-AQA

Violeta Menéndez González, Andrew Gilbert, Graeme Phillipson et al.

In this work, we present an end-to-end network for stereo-consistent image inpainting with the objective of inpainting large missing regions behind objects. The proposed model consists of an edge-guided UNet-like network using Partial Convolutions. We enforce multi-view stereo consistency by introducing a disparity loss. More importantly, we develop a training scheme where the model is learned from realistic stereo masks representing object occlusions, instead of the more common random masks. The technique is trained in a supervised way. Our evaluation shows competitive results compared to previous state-of-the-art techniques.

Sadegh Rahmaniboldaji, Filip Rybansky, Quoc Vuong et al.

Detecting actions in videos, particularly within cluttered scenes, poses significant challenges due to the limitations of 2D frame analysis from a camera perspective. Unlike human vision, which benefits from 3D understanding, recognizing actions in such environments can be difficult. This research introduces a novel approach integrating 3D features and depth maps alongside RGB features to enhance action recognition accuracy. Our method involves processing estimated depth maps through a separate branch from the RGB feature encoder and fusing the features to understand the scene and actions comprehensively. Using the Side4Video framework and VideoMamba, which employ CLIP and VisionMamba for spatial feature extraction, our approach outperformed our implementation of the Side4Video network on the Something-Something V2 dataset. Our code is available at: https://github.com/SadeghRahmaniB/DEAR

Violeta Menéndez González, Andrew Gilbert, Graeme Phillipson et al.

In the field of media production, video editing techniques play a pivotal role. Recent approaches have had great success at performing novel view image synthesis of static scenes. But adding temporal information adds an extra layer of complexity. Previous models have focused on implicitly representing static and dynamic scenes using NeRF. These models achieve impressive results but are costly at training and inference time. They overfit an MLP to describe the scene implicitly as a function of position. This paper proposes ZeST-NeRF, a new approach that can produce temporal NeRFs for new scenes without retraining. We can accurately reconstruct novel views using multi-view synthesis techniques and scene flow-field estimation, trained only with unrelated scenes. We demonstrate how existing state-of-the-art approaches from a range of fields cannot adequately solve this new task and demonstrate the efficacy of our solution. The resulting network improves quantitatively by 15% and produces significantly better visual results.

Edward Fish, Jon Weinbren, Andrew Gilbert

Temporal Action Localization (TAL) aims to identify actions' start, end, and class labels in untrimmed videos. While recent advancements using transformer networks and Feature Pyramid Networks (FPN) have enhanced visual feature recognition in TAL tasks, less progress has been made in the integration of audio features into such frameworks. This paper introduces the Multi-Resolution Audio-Visual Feature Fusion (MRAV-FF), an innovative method to merge audio-visual data across different temporal resolutions. Central to our approach is a hierarchical gated cross-attention mechanism, which discerningly weighs the importance of audio information at diverse temporal scales. Such a technique not only refines the precision of regression boundaries but also bolsters classification confidence. Importantly, MRAV-FF is versatile, making it compatible with existing FPN TAL architectures and offering a significant enhancement in performance when audio data is available.

Xu Dong, Wanqing Li, Anthony Adeyemi-Ejeye et al.

Skeleton-based human action recognition is a powerful approach for understanding human behaviour from pose data, but collecting large-scale, diverse, and well-annotated 3D skeleton datasets is both expensive and labor-intensive. To address this challenge, we propose a conditional generative pipeline for data augmentation in skeleton action recognition. Our method learns the distribution of real skeleton sequences under the constraint of action labels, enabling the synthesis of diverse and high-fidelity data. Even with limited training samples, it can effectively generate skeleton sequences and achieve competitive recognition performance in low-data scenarios, demonstrating strong generalisation in downstream tasks. Specifically, we introduce a Transformer-based encoder-decoder architecture, combined with a generative refinement module and a dropout mechanism, to balance fidelity and diversity during sampling. Experiments on HumanAct12 and the refined NTU-RGBD (NTU-VIBE) dataset show that our approach consistently improves the accuracy of multiple skeleton-based action recognition models, validating its effectiveness in both few-shot and full-data settings. The source code can be found at here.

Cristiana Tiago, Andrew Gilbert, Ahmed S. Beela et al.

Due to privacy issues and limited amount of publicly available labeled datasets in the domain of medical imaging, we propose an image generation pipeline to synthesize 3D echocardiographic images with corresponding ground truth labels, to alleviate the need for data collection and for laborious and error-prone human labeling of images for subsequent Deep Learning (DL) tasks. The proposed method utilizes detailed anatomical segmentations of the heart as ground truth label sources. This initial dataset is combined with a second dataset made up of real 3D echocardiographic images to train a Generative Adversarial Network (GAN) to synthesize realistic 3D cardiovascular Ultrasound images paired with ground truth labels. To generate the synthetic 3D dataset, the trained GAN uses high resolution anatomical models from Computed Tomography (CT) as input. A qualitative analysis of the synthesized images showed that the main structures of the heart are well delineated and closely follow the labels obtained from the anatomical models. To assess the usability of these synthetic images for DL tasks, segmentation algorithms were trained to delineate the left ventricle, left atrium, and myocardium. A quantitative analysis of the 3D segmentations given by the models trained with the synthetic images indicated the potential use of this GAN approach to generate 3D synthetic data, use the data to train DL models for different clinical tasks, and therefore tackle the problem of scarcity of 3D labeled echocardiography datasets.

Soon Yau Cheong, Duygu Ceylan, Armin Mustafa et al.

Recent advancements in diffusion models have significantly enhanced the quality of video generation. However, fine-grained control over camera pose remains a challenge. While U-Net-based models have shown promising results for camera control, transformer-based diffusion models (DiT)-the preferred architecture for large-scale video generation - suffer from severe degradation in camera motion accuracy. In this paper, we investigate the underlying causes of this issue and propose solutions tailored to DiT architectures. Our study reveals that camera control performance depends heavily on the choice of conditioning methods rather than camera pose representations that is commonly believed. To address the persistent motion degradation in DiT, we introduce Camera Motion Guidance (CMG), based on classifier-free guidance, which boosts camera control by over 400%. Additionally, we present a sparse camera control pipeline, significantly simplifying the process of specifying camera poses for long videos. Our method universally applies to both U-Net and DiT models, offering improved camera control for video generation tasks.

Anastasia Anichenko, Frank Guerin, Andrew Gilbert

We investigate a human-like interpretable model of video understanding. Humans recognise complex activities in video by recognising critical spatio-temporal relations among explicitly recognised objects and parts, for example, an object entering the aperture of a container. To mimic this we build on a model which uses positions of objects and hands, and their motions, to recognise the activity taking place. To improve this model we focussed on three of the most confused classes (for this model) and identified that the lack of 3D information was the major problem. To address this we extended our basic model by adding 3D awareness in two ways: (1) A state-of-the-art object detection model was fine-tuned to determine the difference between "Container" and "NotContainer" in order to integrate object shape information into the existing object features. (2) A state-of-the-art depth estimation model was used to extract depth values for individual objects and calculate depth relations to expand the existing relations used our interpretable model. These 3D extensions to our basic model were evaluated on a subset of three superficially similar "Putting" actions from the Something-Something-v2 dataset. The results showed that the container detector did not improve performance, but the addition of depth relations made a significant improvement to performance.

Soon Yau Cheong, Armin Mustafa, Andrew Gilbert

This paper introduces ViscoNet, a novel one-branch-adapter architecture for concurrent spatial and visual conditioning. Our lightweight model requires trainable parameters and dataset size multiple orders of magnitude smaller than the current state-of-the-art IP-Adapter. However, our method successfully preserves the generative power of the frozen text-to-image (T2I) backbone. Notably, it excels in addressing mode collapse, a pervasive issue previously overlooked. Our novel architecture demonstrates outstanding capabilities in achieving a harmonious visual-text balance, unlocking unparalleled versatility in various human image generation tasks, including pose re-targeting, virtual try-on, stylization, person re-identification, and textile transfer.Demo and code are available from project page https://soon-yau.github.io/visconet/ .

Gemma Canet Tarrés, Zhe Lin, Zhifei Zhang et al.

We introduce the first generative model capable of simultaneous multi-object compositing, guided by both text and layout. Our model allows for the addition of multiple objects within a scene, capturing a range of interactions from simple positional relations (e.g., next to, in front of) to complex actions requiring reposing (e.g., hugging, playing guitar). When an interaction implies additional props, like `taking a selfie', our model autonomously generates these supporting objects. By jointly training for compositing and subject-driven generation, also known as customization, we achieve a more balanced integration of textual and visual inputs for text-driven object compositing. As a result, we obtain a versatile model with state-of-the-art performance in both tasks. We further present a data generation pipeline leveraging visual and language models to effortlessly synthesize multimodal, aligned training data.

Adrienne Deganutti, Simon Hadfield, Andrew Gilbert

Audio Description is a narrated commentary designed to aid vision-impaired audiences in perceiving key visual elements in a video. While short-form video understanding has advanced rapidly, a solution for maintaining coherent long-term visual storytelling remains unresolved. Existing methods rely solely on frame-level embeddings, effectively describing object-based content but lacking contextual information across scenes. We introduce DANTE-AD, an enhanced video description model leveraging a dual-vision Transformer-based architecture to address this gap. DANTE-AD sequentially fuses both frame and scene level embeddings to improve long-term contextual understanding. We propose a novel, state-of-the-art method for sequential cross-attention to achieve contextual grounding for fine-grained audio description generation. Evaluated on a broad range of key scenes from well-known movie clips, DANTE-AD outperforms existing methods across traditional NLP metrics and LLM-based evaluations.

Edward Fish, Andrew Gilbert

Few-shot temporal action localization (TAL) methods that adapt large models via single-prompt tuning often fail to produce precise temporal boundaries. This stems from the model learning a non-discriminative mean representation of an action from sparse data, which compromises generalization. We address this by proposing a new paradigm based on multi-prompt ensembles, where a set of diverse, learnable prompts for each action is encouraged to specialize on compositional sub-events. To enforce this specialization, we introduce PLOT-TAL, a framework that leverages Optimal Transport (OT) to find a globally optimal alignment between the prompt ensemble and the video's temporal features. Our method establishes a new state-of-the-art on the challenging few-shot benchmarks of THUMOS'14 and EPIC-Kitchens, without requiring complex meta-learning. The significant performance gains, particularly at high IoU thresholds, validate our hypothesis and demonstrate the superiority of learning distributed, compositional representations for precise temporal localization.

Mona Ahmadian, Amir Shirian, Frank Guerin et al.

Real-world videos often contain overlapping events and complex temporal dependencies, making multimodal interaction modeling particularly challenging. We introduce DEL, a framework for dense semantic action localization, aiming to accurately detect and classify multiple actions at fine-grained temporal resolutions in long untrimmed videos. DEL consists of two key modules: the alignment of audio and visual features that leverage masked self-attention to enhance intra-mode consistency and a multimodal interaction refinement module that models cross-modal dependencies across multiple scales, enabling high-level semantics and fine-grained details. Our method achieves state-of-the-art performance on multiple real-world Temporal Action Localization (TAL) datasets, UnAV-100, THUMOS14, ActivityNet 1.3, and EPIC-Kitchens-100, surpassing previous approaches with notable average mAP gains of +3.3%, +2.6%, +1.2%, +1.7% (verb), and +1.4% (noun), respectively.

Yash Kulthe, Andrew Gilbert, John Collomosse

We present MultiNeRF, a 3D watermarking method that embeds multiple uniquely keyed watermarks within images rendered by a single Neural Radiance Field (NeRF) model, whilst maintaining high visual quality. Our approach extends the TensoRF NeRF model by incorporating a dedicated watermark grid alongside the existing geometry and appearance grids. This extension ensures higher watermark capacity without entangling watermark signals with scene content. We propose a FiLM-based conditional modulation mechanism that dynamically activates watermarks based on input identifiers, allowing multiple independent watermarks to be embedded and extracted without requiring model retraining. MultiNeRF is validated on the NeRF-Synthetic and LLFF datasets, with statistically significant improvements in robust capacity without compromising rendering quality. By generalizing single-watermark NeRF methods into a flexible multi-watermarking framework, MultiNeRF provides a scalable solution for 3D content. attribution.

Kar Balan, Andrew Gilbert, John Collomosse

The rise of Generative AI (GenAI) has sparked significant debate over balancing the interests of creative rightsholders and AI developers. As GenAI models are trained on vast datasets that often include copyrighted material, questions around fair compensation and proper attribution have become increasingly urgent. To address these challenges, this paper proposes a framework called Content ARCs (Authenticity, Rights, Compensation). By combining open standards for provenance and dynamic licensing with data attribution, and decentralized technologies, Content ARCs create a mechanism for managing rights and compensating creators for using their work in AI training. We characterize several nascent works in the AI data licensing space within Content ARCs and identify where challenges remain to fully implement the end-to-end framework.

Mona Ahmadian, Frank Guerin, Andrew Gilbert

This paper demonstrates a self-supervised approach for learning semantic video representations. Recent vision studies show that a masking strategy for vision and natural language supervision has contributed to developing transferable visual pretraining. Our goal is to achieve a more semantic video representation by leveraging the text related to the video content during the pretraining in a fully self-supervised manner. To this end, we present FILS, a novel self-supervised video Feature prediction In semantic Language Space (FILS). The vision model can capture valuable structured information by correctly predicting masked feature semantics in language space. It is learned using a patch-wise video-text contrastive strategy, in which the text representations act as prototypes for transforming vision features into a language space, which are then used as targets for semantically meaningful feature prediction using our masked encoder-decoder structure. FILS demonstrates remarkable transferability on downstream action recognition tasks, achieving state-of-the-art on challenging egocentric datasets, like Epic-Kitchens, Something-SomethingV2, Charades-Ego, and EGTEA, using ViT-Base. Our efficient method requires less computation and smaller batches compared to previous works.

Joseph Chrol-Cannon, Andrew Gilbert, Ranko Lazic et al.

Video activity recognition by deep neural networks is impressive for many classes. However, it falls short of human performance, especially for challenging to discriminate activities. Humans differentiate these complex activities by recognising critical spatio-temporal relations among explicitly recognised objects and parts, for example, an object entering the aperture of a container. Deep neural networks can struggle to learn such critical relationships effectively. Therefore we propose a more human-like approach to activity recognition, which interprets a video in sequential temporal phases and extracts specific relationships among objects and hands in those phases. Random forest classifiers are learnt from these extracted relationships. We apply the method to a challenging subset of the something-something dataset and achieve a more robust performance against neural network baselines on challenging activities.

Dan Ruta, Saeid Motiian, Baldo Faieta et al.

We present ALADIN (All Layer AdaIN); a novel architecture for searching images based on the similarity of their artistic style. Representation learning is critical to visual search, where distance in the learned search embedding reflects image similarity. Learning an embedding that discriminates fine-grained variations in style is hard, due to the difficulty of defining and labelling style. ALADIN takes a weakly supervised approach to learning a representation for fine-grained style similarity of digital artworks, leveraging BAM-FG, a novel large-scale dataset of user generated content groupings gathered from the web. ALADIN sets a new state of the art accuracy for style-based visual search over both coarse labelled style data (BAM) and BAM-FG; a new 2.62 million image dataset of 310,000 fine-grained style groupings also contributed by this work.

Edward Fish, Jon Weinbren, Andrew Gilbert

Movie genre classification is an active research area in machine learning. However, due to the limited labels available, there can be large semantic variations between movies within a single genre definition. We expand these 'coarse' genre labels by identifying 'fine-grained' semantic information within the multi-modal content of movies. By leveraging pre-trained 'expert' networks, we learn the influence of different combinations of modes for multi-label genre classification. Using a contrastive loss, we continue to fine-tune this 'coarse' genre classification network to identify high-level intertextual similarities between the movies across all genre labels. This leads to a more 'fine-grained' and detailed clustering, based on semantic similarities while still retaining some genre information. Our approach is demonstrated on a newly introduced multi-modal 37,866,450 frame, 8,800 movie trailer dataset, MMX-Trailer-20, which includes pre-computed audio, location, motion, and image embeddings.

Kary Ho, Andrew Gilbert, Hailin Jin et al.

We present a neural architecture search (NAS) technique to enhance the performance of unsupervised image de-noising, in-painting and super-resolution under the recently proposed Deep Image Prior (DIP). We show that evolutionary search can automatically optimize the encoder-decoder (E-D) structure and meta-parameters of the DIP network, which serves as a content-specific prior to regularize these single image restoration tasks. Our binary representation encodes the design space for an asymmetric E-D network that typically converges to yield a content-specific DIP within 10-20 generations using a population size of 500. The optimized architectures consistently improve upon the visual quality of classical DIP for a diverse range of photographic and artistic content.

Andrew Gilbert, Marit Holden, Line Eikvil et al.

Two-dimensional echocardiography (2DE) measurements of left ventricle (LV) dimensions are highly significant markers of several cardiovascular diseases. These measurements are often used in clinical care despite suffering from large variability between observers. This variability is due to the challenging nature of accurately finding the correct temporal and spatial location of measurement endpoints in ultrasound images. These images often contain fuzzy boundaries and varying reflection patterns between frames. In this work, we present a convolutional neural network (CNN) based approach to automate 2DE LV measurements. Treating the problem as a landmark detection problem, we propose a modified U-Net CNN architecture to generate heatmaps of likely coordinate locations. To improve the network performance we use anatomically meaningful heatmaps as labels and train with a multi-component loss function. Our network achieves 13.4%, 6%, and 10.8% mean percent error on intraventricular septum (IVS), LV internal dimension (LVID), and LV posterior wall (LVPW) measurements respectively. The design outperforms other networks and matches or approaches intra-analyser expert error.

Andrew Gilbert, Marit Holden, Line Eikvil et al.

Spectral Doppler measurements are an important part of the standard echocardiographic examination. These measurements give important insight into myocardial motion and blood flow providing clinicians with parameters for diagnostic decision making. Many of these measurements can currently be performed automatically with high accuracy, increasing the efficiency of the diagnostic pipeline. However, full automation is not yet available because the user must manually select which measurement should be performed on each image. In this work we develop a convolutional neural network (CNN) to automatically classify cardiac Doppler spectra into measurement classes. We show how the multi-modal information in each spectral Doppler recording can be combined using a meta parameter post-processing mapping scheme and heatmaps to encode coordinate locations. Additionally, we experiment with several state-of-the-art network architectures to examine the tradeoff between accuracy and memory usage for resource-constrained environments. Finally, we propose a confidence metric using the values in the last fully connected layer of the network. We analyze example images that fall outside of our proposed classes to show our confidence metric can prevent many misclassifications. Our algorithm achieves 96% accuracy on a test set drawn from a separate clinical site, indicating that the proposed method is suitable for clinical adoption and enabling a fully automatic pipeline from acquisition to Doppler spectrum measurements.

Andrew Gilbert, Matthew Trumble, Adrian Hilton et al.

We aim to simultaneously estimate the 3D articulated pose and high fidelity volumetric occupancy of human performance, from multiple viewpoint video (MVV) with as few as two views. We use a multi-channel symmetric 3D convolutional encoder-decoder with a dual loss to enforce the learning of a latent embedding that enables inference of skeletal joint positions and a volumetric reconstruction of the performance. The inference is regularised via a prior learned over a dataset of view-ablated multi-view video footage of a wide range of subjects and actions, and show this to generalise well across unseen subjects and actions. We demonstrate improved reconstruction accuracy and lower pose estimation error relative to prior work on two MVV performance capture datasets: Human 3.6M and TotalCapture.

Andrew Gilbert, Marco Volino, John Collomosse et al.

We present a convolutional autoencoder that enables high fidelity volumetric reconstructions of human performance to be captured from multi-view video comprising only a small set of camera views. Our method yields similar end-to-end reconstruction error to that of a probabilistic visual hull computed using significantly more (double or more) viewpoints. We use a deep prior implicitly learned by the autoencoder trained over a dataset of view-ablated multi-view video footage of a wide range of subjects and actions. This opens up the possibility of high-end volumetric performance capture in on-set and prosumer scenarios where time or cost prohibit a high witness camera count.

Matthew Trumble, Andrew Gilbert, Adrian Hilton et al.

We present a method for simultaneously estimating 3D human pose and body shape from a sparse set of wide-baseline camera views. We train a symmetric convolutional autoencoder with a dual loss that enforces learning of a latent representation that encodes skeletal joint positions, and at the same time learns a deep representation of volumetric body shape. We harness the latter to up-scale input volumetric data by a factor of $4 \times$, whilst recovering a 3D estimate of joint positions with equal or greater accuracy than the state of the art. Inference runs in real-time (25 fps) and has the potential for passive human behaviour monitoring where there is a requirement for high fidelity estimation of human body shape and pose.

Andrew Gilbert, Richard Bowden

Within the field of image and video recognition, the traditional approach is a dataset split into fixed training and test partitions. However, the labelling of the training set is time-consuming, especially as datasets grow in size and complexity. Furthermore, this approach is not applicable to the home user, who wants to intuitively group their media without tirelessly labelling the content. Our interactive approach is able to iteratively cluster classes of images and video. Our approach is based around the concept of an image signature which, unlike a standard bag of words model, can express co-occurrence statistics as well as symbol frequency. We efficiently compute metric distances between signatures despite their inherent high dimensionality and provide discriminative feature selection, to allow common and distinctive elements to be identified from a small set of user labelled examples. These elements are then accentuated in the image signature to increase similarity between examples and pull correct classes together. By repeating this process in an online learning framework, the accuracy of similarity increases dramatically despite labelling only a few training examples. To demonstrate that the approach is agnostic to media type and features used, we evaluate on three image datasets (15 scene, Caltech101 and FG-NET), a mixed text and image dataset (ImageTag), a dataset used in active learning (Iris) and on three action recognition datasets (UCF11, KTH and Hollywood2). On the UCF11 video dataset, the accuracy is 86.7% despite using only 90 labelled examples from a dataset of over 1200 videos, instead of the standard 1122 training videos. The approach is both scalable and efficient, with a single iteration over the full UCF11 dataset of around 1200 videos taking approximately 1 minute on a standard desktop machine.