Virtual Category Learning: A Semi-Supervised Learning Method for Dense Prediction With Extremely Limited Labels.

Due to the costliness of labelled data in real-world applications, semi-supervised learning, underpinned by pseudo labelling, is an appealing solution. However, handling confusing samples is nontrivial: discarding valuable confusing samples would compromise the model generalisation while using them for training would exacerbate the issue of confirmation bias caused by the resulting inevitable mislabelling. To solve this problem, this paper proposes to use confusing samples proactively without label correction.

Semi-Supervised Unpaired Medical Image Segmentation Through Task-Affinity Consistency.

Deep learning-based semi-supervised learning (SSL) algorithms are promising in reducing the cost of manual annotation of clinicians by using unlabelled data, when developing medical image segmentation tools. However, to date, most existing semi-supervised learning (SSL) algorithms treat the labelled images and unlabelled images separately and ignore the explicit connection between them; this disregards essential shared information and thus hinders further performance improvements. To mine the shared information between the labelled and unlabelled images, we introduce a class-specific representation extraction approach, in which a task-affinity module is specifically designed for representation extraction.

Using virtual environments to facilitate refugee integration in third countries.

Virtual experiences (VEs) have significant potential to enrich emotional interactions, to encourage socialisation and improve communication. In education, VEs offer new approaches for delivering content. In this paper we consider the application of VEs for assisting refugees in Senegal to learn how to navigate the complexities of the UK health system; a substantial stumbling block for their integration into society and for their own health.

Deep Synthesis of Cloud Lighting.

Current appearance models for the sky are able to represent clear-sky illumination to a high degree of accuracy. However, these models all lack a common feature of real skies: clouds. These are an essential component for many applications which rely on realistic skies, such as image editing and synthesis.

Region-Object Relation-Aware Dense Captioning via Transformer.

Dense captioning provides detailed captions of complex visual scenes. While a number of successes have been achieved in recent years, there are still two broad limitations: 1) most existing methods adopt an encoder-decoder framework, where the contextual information is sequentially encoded using long short-term memory (LSTM). However, the forget gate mechanism of LSTM makes it vulnerable when dealing with a long sequence and 2) the vast majority of prior arts consider regions of interests (RoIs) equally important, thus failing to focus on more informative regions.

Deep HDR Hallucination for Inverse Tone Mapping.

Inverse Tone Mapping (ITM) methods attempt to reconstruct High Dynamic Range (HDR) information from Low Dynamic Range (LDR) image content. The dynamic range of well-exposed areas must be expanded and any missing information due to over/under-exposure must be recovered (hallucinated). The majority of methods focus on the former and are relatively successful, while most attempts on the latter are not of sufficient quality, even ones based on Convolutional Neural Networks (CNNs).

Cloud-Based Dynamic GI for Shared VR Experiences.

Real-time high-fidelity rendering requires the use of expensive high-end hardware, even when rendering moderately complex scenes. Interactive streaming services and cloud gaming have somewhat mitigated the problem at the cost of response lag. In this article, we present Regular Grid Global Illumination (ReGGI), a distributed rendering pipeline that eliminates response lag and provides cloud-based dynamic GI for low-powered devices such as smartphones and the class of devices typically used in untethered VR headsets.

A Machine-Learning-Driven Sky Model.

Sky illumination is responsible for much of the lighting in a virtual environment. A machine-learning-based approach can compactly represent sky illumination from both existing analytic sky models and from captured environment maps. The proposed approach can approximate the captured lighting at a significantly reduced memory cost and enable smooth transitions of sky lighting to be created from a small set of environment maps captured at discrete times of day.

Importance Driven Environment Map Sampling.

In this paper we present an efficient method for supporting image based lighting (IBL) for bidirectional methods. This improves both sampling of the environment, and the detection and sampling of important regions of the scene, such as windows and doors. These parts of the scene often have a small area proportional to that of the entire scene, so paths which pass through them are generated with a low probability.

High-fidelity interactive rendering on desktop grids.

Desktop grids combine arbitrary computational resources connected to a network. However, the prevalent interactive rendering algorithms can't seamlessly handle the variable computational power offered by a desktop grid's nondedicated resources. In this article, a method for achieving interactive high-fidelity rendering on nondedicated machines such as desktop grids is developed, without the expensive requirements of a dedicated render farm.