Shunian Chen, Xinyuan Xie, Zheshu Chen, Owen Lee, Liyan Zhao, Zhan Su, Qilin Sun, Benyou Wang. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Capturing the full plenoptic light transport across spatial, angular, and temporal dimensions has long been a pursuit in computational imaging, yet it remains fundamentally constrained by the high dimensionality of the sampling space and the physical inaccessibility of scene regions due to self-occlusions. While time-resolved imaging records the temporal axis, existing methods are bottlenecked by the combinatorial complexity of the plenoptic function. This high dimensionality makes dense omni-dimensional sampling physically prohibitive. Simultaneously, tight coupling between illumination and viewpoint in current systems also precludes the full acquisition of plenoptic light transport. In this work, we present GenPIE, a Generative Plenoptic Imager designed to bridge the gap between sparse physical observations and high-dimensional light transport. We introduce a decoupled laser-detector hardware setup that enables independent control over illumination and detection, allowing for active probing of indirect light paths. To overcome the ill-posedness of sparse sampling and physical blind spots, we propose a generative inverse transient rendering framework. Our approach leverages 3D foundation models to provide strong semantic and 3D geometric priors for initialization, which are subsequently refined through a differentiable transient path tracer to ensure physically grounded adherence to the Transient Rendering Equation. We demonstrate that GenPIE supports a range of applications that are challenging for steady-state or purely neural methods, including disentangling multi-bounce light transport directly from captured transient videos, time unwarping, and time-resolved relighting. The project page is at https://wangzh1.github.io/GenPIE.

High dynamic range (HDR) images, with their rich tone and detail reproduction, hold significant potential to enhance computer vision systems, particularly in autonomous driving. However, most neural networks for embedded vision are trained on low dynamic range (LDR) inputs and suffer substantial performance degradation when handling high-bit-depth HDR images due to the challenges posed by extreme dynamic ranges. In this paper, we propose a novel tone mapping method that not only bridges the gap between HDR RAW inputs and the LDR sRGB requirements of detection networks but also achieves end-to-end optimization with the downstream tasks. Instead of relying on traditional image signal processing (ISP) pipeline, we introduce neural photometric calibration to regularize dynamic ranges and a scaling-invariant local tone mapping module to preserve image details. In addition, our architecture also supports performance transfer finetuning, enabling efficient adaptation from the LDR model to the HDR RAW model with minimal cost. The proposed method outperforms traditional tone mapping algorithms and advanced AI-ISP methods in challenging automotive HDR scenes. Moreover, our pipeline achieves real-time processing of 4K high-bit-depth HDR inputs on the Nvidia Jetson platform.

Most well-established and widely used color difference (CD) metrics are handcrafted and subjectcalibrated against uniformly colored patches, which do not generalize well to photographic images characterized by natural scene complexities. Constructing CD formulae for photo-graphic images is still an active research topic in imaging/illumination, vision science, and color science communities. In this paper, we aim to learn a deep CD metric for photographic images with four desirable properties. First, it well aligns with the observations in vision science that color and form are linked inextricably in visual cortical processing. Second, it is a proper metric in the mathematical sense. Third, it computes accurate CDs between photographic images, differing mainly in color appearances. Fourth, it is robust to mild geometric distortions (e.g., translation or due to parallax), which are often present in photographic images of the same scene captured by different digital cameras. We show that all four properties can be satisfied at once by learning a multi-scale autoregressive normalizing flow for feature transform, followed by the Euclidean distance which is linearly proportional to the human perceptual CD. Quantitative and qualitative experiments on the large-scale SPCD dataset demonstrate the promise of the learned CD metric. Source code is available at https://github.com/haoychen3/CD-Flow.

Immersive user experiences in live VR/AR performances require a fast and accurate free-view rendering of the performers. Existing methods are mainly based on Pixel-aligned Implicit Functions (PIFu) or Neural Radiance Fields (NeRF). However, while PIFu-based methods usually fail to produce photorealistic view-dependent textures, NeRF-based methods typically lack local geometry accuracy and are computationally heavy ( e.g. , dense sampling of 3D points, additional fine-tuning, or pose estimation). In this work, we propose a novel generalizable method, named SAILOR, to create high-quality human free-view videos from very sparse RGBD live streams. To produce view-dependent textures while preserving locally accurate geometry, we integrate PIFu and NeRF such that they work synergistically by conditioning the PIFu on depth and then rendering view-dependent textures through NeRF. Specifically, we propose a novel network, named SRONet, for this hybrid representation. SRONet can handle unseen performers without fine-tuning. Besides, a neural blending-based ray interpolation approach, a tree-based voxel-denoising scheme, and a parallel computing pipeline are incorporated to reconstruct and render live free-view videos at 10 fps on average. To evaluate the rendering performance, we construct a real-captured RGBD benchmark from 40 performers. Experimental results show that SAILOR outperforms existing human reconstruction and performance capture methods.

Hyperspectral imaging is an essential imaging modality for a wide range of applications, especially in remote sensing, agriculture, and medicine. Inspired by existing hyperspectral cameras that are either slow, expensive, or bulky, reconstructing hyperspectral images (HSIs) from a low-budget snapshot measurement has drawn wide attention. By mapping a truncated numerical optimization algorithm into a network with a fixed number of phases, recent deep unfolding networks (DUNs) for spectral snapshot compressive sensing (SCI) have achieved remarkable success. However, DUNs are far from reaching the scope of industrial applications limited by the lack of cross-phase feature interaction and adaptive parameter adjustment. In this paper, we propose a novel Hyperspectral Explicable Reconstruction and Optimal Sampling deep Network for SCI, dubbed HerosNet, which includes several phases under the ISTA-unfolding framework. Each phase can flexibly simulate the sensing matrix and contextually adjust the step size in the gradient descent step, and hierarchically fuse and interact the hidden states of previous phases to effectively recover current HSI frames in the proximal mapping step. Simultaneously, a hardware-friendly optimal binary mask is learned end-to-end to further improve the reconstruction performance. Finally, our HerosNet is validated to outperform the state-of-the-art methods on both simulation and real datasets by large margins. The source code is available at https://github.com/jianzhangcs/HerosNet.

Imaging systems have long been designed in separated steps: experience-driven optical design followed by sophisticated image processing. Although recent advances in computational imaging aim to bridge the gap in an end-to-end fashion, the image formation models used in these approaches have been quite simplistic, built either on simple wave optics models such as Fourier transform, or on similar paraxial models. Such models only support the optimization of a single lens surface, which limits the achievable image quality. To overcome these challenges, we propose a general end-to-end complex lens design framework enabled by a differentiable ray tracing image formation model. Specifically, our model relies on the differentiable ray tracing rendering engine to render optical images in the full field by taking into account all on/off-axis aberrations governed by the theory of geometric optics. Our design pipeline can jointly optimize the lens module and the image reconstruction network for a specific imaging task. We demonstrate the effectiveness of the proposed method on two typical applications, including large field-of-view imaging and extended depth-of-field imaging. Both simulation and experimental results show superior image quality compared with conventional lens designs. Our framework offers a competitive alternative for the design of modern imaging systems.

High-dynamic range (HDR) imaging is an essential imaging modality for a wide range of applications in uncontrolled environments, including autonomous driving, robotics, and mobile phone cameras. However, existing HDR techniques in commodity devices struggle with dynamic scenes due to multi-shot acquisition and post-processing time, e.g. mobile phone burst photography, making such approaches unsuitable for real-time applications. In this work, we propose a method for snapshot HDR imaging by learning an optical HDR encoding in a single image which maps saturated highlights into neighboring unsaturated areas using a diffractive optical element (DOE). We propose a novel rank-1 parameterization of the proposed DOE which avoids vast trainable parameters and keeps high frequencies' encoding compared with conventional end-to-end design methods. We further propose a reconstruction network tailored to this rank-1 parametrization for recovery of clipped information from the encoded measurements. The proposed end-to-end framework is validated through simulation and real-world experiments and improves the PSNR by more than 7 dB over state-of-the-art end-to-end designs.

Single Photon Avalanche Photodiodes (SPADs) have recently received a lot of attention in imaging and vision applications due to their excellent performance in low-light conditions, as well as their ultra-high temporal resolution. Unfortunately, like many evolving sensor technologies, image sensors built around SPAD technology currently suffer from a low pixel count. In this work, we investigate a simple, low-cost, and compact optical coding camera design that supports high-resolution image reconstructions from raw measurements with low pixel counts. We demonstrate this approach for regular intensity imaging, depth imaging, as well transient imaging. Our method uses an end-to-end framework to simultaneously optimize the optical design and a reconstruction network for obtaining super-resolved images from raw measurements. The optical design space is that of an engineered point spread function (implemented with diffractive optics), which can be considered an optimized anti-aliasing filter to preserve as much high-resolution information as possible despite imaging with a low pixel count, low fill-factor SPAD array. We further investigate a deep network for reconstruction. The effectiveness of this joint design and reconstruction approach is demonstrated for a range of different applications, including high-speed imaging, and time of flight depth imaging, as well as transient imaging. While our work specifically focuses on low-resolution SPAD sensors, similar approaches should prove effective for other emerging image sensor technologies with low pixel counts and low fill-factors.

Typical camera optics consist of a system of individual elements that are designed to compensate for the aberrations of a single lens. Recent computational cameras shift some of this correction task from the optics to post-capture processing, reducing the imaging optics to only a few optical elements. However, these systems only achieve reasonable image quality by limiting the field of view (FOV) to a few degrees - effectively ignoring severe off-axis aberrations with blur sizes of multiple hundred pixels. In this paper, we propose a lens design and learned reconstruction architecture that lift this limitation and provide an order of magnitude increase in field of view using only a single thin-plate lens element. Specifically, we design a lens to produce spatially shift-invariant point spread functions, over the full FOV, that are tailored to the proposed reconstruction architecture. We achieve this with a mixture PSF, consisting of a peak and and a low-pass component, which provides residual contrast instead of a small spot size as in traditional lens designs. To perform the reconstruction, we train a deep network on captured data from a display lab setup, eliminating the need for manual acquisition of training data in the field. We assess the proposed method in simulation and experimentally with a prototype camera system. We compare our system against existing single-element designs, including an aspherical lens and a pinhole, and we compare against a complex multielement lens, validating high-quality large field-of-view (i.e. 53°) imaging performance using only a single thin-plate element.

Time-of-flight depth imaging and transient imaging are two imaging modalities that have recently received a lot of interest. Despite much research, existing hardware systems are limited either in terms of temporal resolution or are prohibitively expensive. Arrays of Single Photon Avalanche Diodes (SPADs) promise to fill this gap by providing higher temporal resolution at an affordable cost. Unfortunately SPAD arrays are to date only available in relatively small resolutions. In this work we aim to overcome the spatial resolution limit of SPAD arrays by employing a compressive sensing camera design. Using a DMD and custom optics, we achieve an image resolution of up to 800×400 on SPAD Arrays of resolution 64×32. Using our new data fitting model for the time histograms, we suppress the noise while abstracting the phase and amplitude information, so as to realize a temporal resolution of a few tens of picoseconds.

Depth-dependent defocus results in a limited depth-of-field in consumer-level cameras. Computational imaging provides alternative solutions to resolve all-in-focus images with the assistance of designed optics and algorithms. In this work, we extend the concept of focal sweep from refractive optics to diffractive optics, where we fuse multiple focal powers onto one single element. In contrast to state-of-the-art sweep models, ours can generate better-conditioned point spread function (PSF) distributions along the expected depth range with drastically shortened (40%) sweep distance. Further by encoding axially asymmetric PSFs subject to color channels, and then sharing sharp information across channels, we preserve details as well as color fidelity. We prototype two diffractive imaging systems that work in the monochromatic and RGB color domain. Experimental results indicate that the depth-of-field can be significantly extended with fewer artifacts remaining after the deconvolution.

Computational caustics and light steering displays offer a wide range of interesting applications, ranging from art works and architectural installations to energy efficient HDR projection. In this work we expand on this concept by encoding several target images into pairs of front and rear phase-distorting surfaces. Different target holograms can be decoded by mixing and matching different front and rear surfaces under specific geometric alignments. Our approach, which we call mix-and-match holography, is made possible by moving from a refractive caustic image formation process to a diffractive, holographic one. This provides the extra bandwidth that is required to multiplex several images into pairing surfaces. We derive a detailed image formation model for the setting of holographic projection displays, as well as a multiplexing method based on a combination of phase retrieval methods and complex matrix factorization. We demonstrate several application scenarios in both simulation and physical prototypes.