Training Neural Networks on RAW and HDR Images for Restoration Tasks

The vast majority of standard image and video content available online is represented in display-encoded color spaces, in which pixel values are conveniently scaled to a limited range (0-1) and the color distribution is approximately perceptually uniform. In contrast, both camera RAW and high dynamic range (HDR) images are often represented in linear color spaces, in which color values are linearly related to colorimetric quantities of light. While training on commonly available display-encoded images is a well-established practice, there is no consensus on how neural networks should be trained for tasks on RAW and HDR images in linear color spaces. In this work, we test several approaches on three popular image restoration applications: denoising, deblurring, and single-image super-resolution. We examine whether HDR/RAW images need to be display-encoded using popular transfer functions (PQ, PU21, mu-law), or whether it is better to train in linear color spaces, but use loss functions that correct for perceptual non-uniformity. Our results indicate that neural networks train significantly better on HDR and RAW images represented in display-encoded color spaces, which offer better perceptual uniformity than linear spaces. This small change to the training strategy can bring a very substantial gain in performance, up to 10-15 dB.

Stereoscopic Depth Perception Through Foliage

Both humans and computational methods struggle to discriminate the depths of objects hidden beneath foliage. However, such discrimination becomes feasible when we combine computational optical synthetic aperture sensing with the human ability to fuse stereoscopic images. For object identification tasks, as required in search and rescue, wildlife observation, surveillance, and early wildfire detection, depth assists in differentiating true from false findings, such as people, animals, or vehicles vs. sun-heated patches at the ground level or in the tree crowns, or ground fires vs. tree trunks. We used video captured by a drone above dense woodland to test users' ability to discriminate depth. We found that this is impossible when viewing monoscopic video and relying on motion parallax. The same was true with stereoscopic video because of the occlusions caused by foliage. However, when synthetic aperture sensing was used to reduce occlusions and disparity-scaled stereoscopic video was presented, whereas computational (stereoscopic matching) methods were unsuccessful, human observers successfully discriminated depth. This shows the potential of systems which exploit the synergy between computational methods and human vision to perform tasks that neither can perform alone.

Perceptual Quality Assessment of NeRF and Neural View Synthesis Methods for Front-Facing Views

Neural view synthesis (NVS) is one of the most successful techniques for synthesizing free viewpoint videos, capable of achieving high fidelity from only a sparse set of captured images. This success has led to many variants of the techniques, each evaluated on a set of test views typically using image quality metrics such as PSNR, SSIM, or LPIPS. There has been a lack of research on how NVS methods perform with respect to perceived video quality. We present the first study on perceptual evaluation of NVS and NeRF variants. For this study, we collected two datasets of scenes captured in a controlled lab environment as well as in-the-wild. In contrast to existing datasets, these scenes come with reference video sequences, allowing us to test for temporal artifacts and subtle distortions that are easily overlooked when viewing only static images. We measured the quality of videos synthesized by several NVS methods in a well-controlled perceptual quality assessment experiment as well as with many existing state-of-the-art image/video quality metrics. We present a detailed analysis of the results and recommendations for dataset and metric selection for NVS evaluation.

G-SemTMO: Tone Mapping with a Trainable Semantic Graph

A Tone Mapping Operator (TMO) is required to render images with a High Dynamic Range (HDR) on media with limited dynamic capabilities. TMOs compress the dynamic range with the aim of preserving the visually perceptual cues of the scene. Previous literature has established the benefits of TMOs being semantic aware, understanding the content in the scene to preserve the cues better. Expert photographers analyze the semantic and the contextual information of a scene and decide tonal transformations or local luminance adjustments. This process can be considered a manual analogy to tone mapping. In this work, we draw inspiration from an expert photographer's approach and present a Graph-based Semantic-aware Tone Mapping Operator, G-SemTMO. We leverage semantic information as well as the contextual information of the scene in the form of a graph capturing the spatial arrangements of its semantic segments. Using Graph Convolutional Network (GCN), we predict intermediate parameters called Semantic Hints and use these parameters to apply tonal adjustments locally to different semantic segments in the image. In addition, we also introduce LocHDR, a dataset of 781 HDR images tone mapped manually by an expert photo-retoucher with local tonal enhancements. We conduct ablation studies to show that our approach, G-SemTMO\footnote{Code and dataset to be published with the final version of the manuscript}, can learn both global and local tonal transformations from a pair of input linear and manually retouched images by leveraging the semantic graphs and produce better results than both classical and learning based TMOs. We also conduct ablation experiments to validate the advantage of using GCN.

Active Sampling for Pairwise Comparisons via Approximate Message Passing and Information Gain Maximization

Pairwise comparison data arise in many domains with subjective assessment experiments, for example in image and video quality assessment. In these experiments observers are asked to express a preference between two conditions. However, many pairwise comparison protocols require a large number of comparisons to infer accurate scores, which may be unfeasible when each comparison is time-consuming (e.g. videos) or expensive (e.g. medical imaging). This motivates the use of an active sampling algorithm that chooses only the most informative pairs for comparison. In this paper we propose ASAP, an active sampling algorithm based on approximate message passing and expected information gain maximization. Unlike most existing methods, which rely on partial updates of the posterior distribution, we are able to perform full updates and therefore much improve the accuracy of the inferred scores. The algorithm relies on three techniques for reducing computational cost: inference based on approximate message passing, selective evaluations of the information gain, and selecting pairs in a batch that forms a minimum spanning tree of the inverse of information gain. We demonstrate, with real and synthetic data, that ASAP offers the highest accuracy of inferred scores compared to the existing methods. We also provide an open-source GPU implementation of ASAP for large-scale experiments.

Exploiting Synthetically Generated Data with Semi-Supervised Learning for Small and Imbalanced Datasets

Data augmentation is rapidly gaining attention in machine learning. Synthetic data can be generated by simple transformations or through the data distribution. In the latter case, the main challenge is to estimate the label associated to new synthetic patterns. This paper studies the effect of generating synthetic data by convex combination of patterns and the use of these as unsupervised information in a semi-supervised learning framework with support vector machines, avoiding thus the need to label synthetic examples. We perform experiments on a total of 53 binary classification datasets. Our results show that this type of data over-sampling supports the well-known cluster assumption in semi-supervised learning, showing outstanding results for small high-dimensional datasets and imbalanced learning problems.