Lensless cameras, innovatively replacing traditional lenses for ultra-thin, flat optics, encode light directly onto sensors, producing images that are not immediately recognizable. This compact, lightweight, and cost-effective imaging solution offers inherent privacy advantages, making it attractive for privacy-sensitive applications like face verification. Typical lensless face verification adopts a two-stage process of reconstruction followed by verification, incurring privacy risks from reconstructed faces and high computational costs. This paper presents an end-to-end optimization approach for privacy-preserving face verification directly on encoded lensless captures, ensuring that the entire software pipeline remains encoded with no visible faces as intermediate results. To achieve this, we propose several techniques to address unique challenges from the lensless setup which precludes traditional face detection and alignment. Specifically, we propose a face center alignment scheme, an augmentation curriculum to build robustness against variations, and a knowledge distillation method to smooth optimization and enhance performance. Evaluations in both simulation and real environments demonstrate that our method outperforms two-stage lensless verification while enhancing privacy and efficiency.

Accurate blur estimation is essential for high-performance imaging in various applications. Blur is typically represented by the point spread function (PSF). In this paper, we propose a physics-informed PSF learning framework for imaging systems, consisting of a simple calibration followed by a learning process. Our framework could achieve both high accuracy and universal applicability. Inspired by the Seidel PSF model for representing spatially varying PSF, we identify its limitations in optimization and introduce a novel wavefront-based PSF model accompanied by an optimization strategy, both reducing optimization complexity and improving estimation accuracy. Moreover, our wavefront-based PSF model is independent of lens parameters, eliminating the need for prior knowledge of the lens. To validate our approach, we compare it with recent PSF estimation methods (Degradation Transfer and Fast Two-step) through a deblurring task, where all the estimated PSFs are used to train state-of-the-art deblurring algorithms. Our approach demonstrates improvements in image quality in simulation and also showcases noticeable visual quality improvements on real captured images. The code and models are public.

Designing 3D scenes is traditionally a challenging and laborious task that demands both artistic expertise and proficiency with complex software. Recent advances in text-to-3D generation have greatly simplified this process by letting users create scenes based on simple text descriptions. However, as these methods generally require extra training or in-context learning, their performance is often hindered by the limited availability of high-quality 3D data. In contrast, modern text-to-image models learned from web-scale images can generate scenes with diverse, reliable spatial layouts and consistent, visually appealing styles. Our key insight is that instead of learning directly from 3D scenes, we can leverage generated 2D images as an intermediary to guide 3D synthesis. In light of this, we introduce ArtiScene, a training-free automated pipeline for scene design that integrates the flexibility of free-form text-to-image generation with the diversity and reliability of 2D intermediary layouts. First, we generate 2D images from a scene description, then extract the shape and appearance of objects to create 3D models. These models are assembled into the final scene using geometry, position, and pose information derived from the same intermediary image. Being generalizable to a wide range of scenes and styles, ArtiScene outperforms state-of-the-art benchmarks by a large margin in layout and aesthetic quality by quantitative metrics. It also averages a 74.89 % winning rate in extensive user studies and 95.07 % in GPT-4o evaluation.

Recent advancements in generative models have significantly improved novel view synthesis (NVS) from multi-view data. However, existing methods depend on external multiview alignment processes, such as explicit pose estimation or pre-reconstruction, which limits their flexibility and accessibility, especially when alignment is unstable due to insufficient overlap or occlusions between views. In this paper, we propose NVComposer, a novel approach that eliminates the need for explicit external alignment. NVComposer enables the generative model to implicitly infer spatial and geometric relationships between multiple conditional views by introducing two key components: 1) an image-pose dualstream diffusion model that simultaneously generates target novel views and condition camera poses, and 2) a geometry-aware feature alignment module that distills geometric priors from dense stereo models during training. Extensive experiments demonstrate that NVComposer achieves state-of-the-art performance in generative multi-view NVS tasks, removing the reliance on external alignment and thus improving model accessibility. Our approach shows substantial improvements in synthesis quality as the number of unposed input views increases, highlighting its potential for more flexible and accessible generative NVS systems.

We present the Generalized Spatial Propagation Network (GSPN), a new attention mechanism optimized for vision tasks that inherently captures 2D spatial structures. Existing attention models, including transformers, linear attention, and state-space models like Mamba, process multidimensional data as 1D sequences, compromising spatial coherence and efficiency. GSPN overcomes these limitations by directly operating on spatially coherent image data and forming dense pairwise connections through a line-scan approach. Central to GSPN is the Stability-Context Condition, which ensures stable, long-context propagation across 2D sequences and reduces the effective sequence length to $\sqrt N $ for a square map with N elements, which significantly enhances computational efficiency. With learnable, input-dependent weights and no reliance on positional embeddings, GSPN achieves superior spatial fidelity and state-of-the-art performance in vision tasks, including ImageNet classification, class-guided image generation, and text-to-image generation. Notably, GSPN accelerates SD-XL with softmax-attention by over 84× when generating 16K images. Project page: https://whj363636.github.io/GSPN/

Reflection removal is challenging due to complex light interactions, where reflections obscure important details and hinder scene understanding. Polarization naturally provides a powerful cue to distinguish between reflected and transmitted light, enabling more accurate reflection removal. However, existing methods often rely on small-scale or synthetic datasets, which fail to capture the diversity and complexity of real-world scenarios. To this end, we construct a large-scale dataset, PolaRGB, for Polarization-based reflection removal of RGB images, which enables us to train models that generalize effectively across a wide range of real-world scenarios. The PolaRGB dataset contains 6,500 well-aligned mixed-transmission image pairs, 8× larger than existing polarization datasets, and is the first to include both RGB and polarization images captured across diverse indoor and outdoor environments with varying lighting conditions. Besides, to fully exploit the potential of polarization cues for reflection removal, we introduce PolarFree, which leverages diffusion process to generate reflection-free cues for accurate reflection removal. Extensive experiments show that PolarFree significantly enhances image clarity in challenging reflective scenarios, setting a new benchmark for polarized imaging and reflection removal. Code and dataset are available at https://github.com/mdyao/PolarFree.

Image datasets with paired mean opinion scores (MOS) enable the quantization of the perceptual differences between images and are of great significance for images taken underwater. In our previous research, a pairwise label underwater image quality subjective ranking (PLUIQR) method was proposed. In this paper, we take a further step by designing a post-reliability verification for the PLUIQR and releasing a publicly accessible underwater image quality dataset called PCUID. For the raw paired voting, steps including a triangular cycle error (TCE) criterion and reprocessing of the dispute data are performed. Meanwhile, the group maximum differentiation (gMAD) method is used to evaluate the performance of underwater quality assessment methods (UIQA) on images with similar quality. That illustrates the proposed dataset enabling UIQA methods to their judgment toward subtle quality differences. The dataset is available at https://github.com/JOU-UIP/PCUID.

Efficient vision transformer remains a bottleneck for high-resolution images and long-video related real-world applications. Generalized Spatial Propagation Network (GSPN) \cite{wang2025parallel} addresses this by replacing quadratic self-attention with a line-scan propagation scheme, bringing the cost close to linear in the number of rows or columns, while retaining accuracy. Despite this advancement, the existing GSPN implementation still suffers from (i) heavy overhead due to repeatedly launching GPU kernels, (ii) excessive data transfers from global GPU memory, and (iii) redundant computations caused by maintaining separate propagation weights for each channel. We introduce GSPN-2, a joint algorithm–system redesign. In particular, we eliminate thousands of micro-launches from the previous implementation into one single 2D kernel, explicitly pin one warp to each channel slice, and stage the previous column's activations in shared memory. On the model side, we introduce a set of channel-shared propagation weights that replace per-channel matrices, trimming parameters, and align naturally with the affinity map used in transformer attention. Experiments demonstrate GSPN-2's effectiveness across image classification and text-to-image synthesis tasks, matching transformer-level accuracy with significantly lower computational cost. GSPN-2 establishes a new efficiency frontier for modeling global spatial context in vision applications through its unique combination of structured matrix transformations and GPU-optimized implementation.

Reconstructing fast-dynamic scenes from multi-view videos is crucial for high-speed motion analysis and realistic 4D reconstruction. However, the majority of 4D capture systems are limited to frame rates below 30 FPS (frames per second), and a direct 4D reconstruction of high-speed motion from low FPS input may lead to undesirable results. In this work, we propose a high-speed 4D capturing system only using low FPS cameras, through novel capturing and processing modules. On the capturing side, we propose an asynchronous capture scheme that increases the effective frame rate by staggering the start times of cameras. By grouping cameras and leveraging a base frame rate of 25 FPS, our method achieves an equivalent frame rate of 100–200 FPS without requiring specialized high-speed cameras. On processing side, we also propose a novel generative model to fix artifacts caused by 4D sparse-view reconstruction, as asynchrony reduces the number of viewpoints at each timestamp. Specifically, we propose to train a video-diffusion-based artifact-fix model for sparse 4D reconstruction, which refines missing details, maintains temporal consistency, and improves overall reconstruction quality. Experimental results demonstrate that our method significantly enhances high-speed 4D reconstruction compared to synchronous capture. Project page: https://openimaginglab.github.io/4DSloMo/

Modern end-to-end image signal processors (ISPs) can learn complex mappings from RAW/XYZ data to sRGB (and vice versa), opening new possibilities in image processing. However, the growing diversity of camera models, particularly in mobile devices, renders the development of individual ISPs unsustainable due to their limited versatility and adaptability across varied camera systems. In this paper, we introduce Uni-ISP, a novel pipeline that unifies ISP learning for diverse mobile cameras, delivering a highly accurate and adaptable processor. The core of Uni-ISP is leveraging device-aware embeddings through learning forward/inverse ISPs and its special training scheme. By doing so, Uni-ISP not only improves the performance of forward and inverse ISPs but also unlocks new applications previously inaccessible to conventional learned ISPs. To support this work, we construct a real-world 4K dataset, FiveCam, comprising more than 2,400 pairs of sRGB-RAW images captured synchronously by five smartphone cameras. Extensive experiments validate Uni-ISP's accuracy in learning forward and inverse ISPs (with improvements of +2.4dB/1.5dB PSNR), versatility in enabling new applications, and adaptability to new camera models.

This work introduces a new Transformer model called Cached Transformer, which uses Gated Recurrent Cached (GRC) attention to extend the self-attention mechanism with a differentiable memory cache of tokens. GRC attention enables attending to both past and current tokens, increasing the receptive field of attention and allowing for exploring long-range dependencies. By utilizing a recurrent gating unit to continuously update the cache, our model achieves significant advancements in \textbf{six} language and vision tasks, including language modeling, machine translation, ListOPs, image classification, object detection, and instance segmentation. Furthermore, our approach surpasses previous memory-based techniques in tasks such as language modeling and displays the ability to be applied to a broader range of situations.

Reconstructing High Dynamic Range (HDR) video from image sequences captured with alternating exposures is challenging, especially in the presence of large camera or object motion. Existing methods typically align low dynamic range sequences using optical flow or attention mechanism for deghosting. However, they often struggle to handle large complex motions and are computation-ally expensive. To address these challenges, we propose a robust and efficient flow estimator tailored for real-time HDR video reconstruction, named HDRFlow. HDRFlow has three novel designs: an HDR-domain alignment loss (HALoss), an efficient flow network with a multi-size large kernel (MLK), and a new HDR flow training scheme. The HALoss supervises our flow network to learn an HDR-oriented flow for accurate alignment in saturated and dark regions. The MLK can effectively model large motions at a negligible cost. In addition, we incorporate synthetic data, Sintel, into our training dataset, utilizing both its provided forward flow and backward flow generated by us to super-vise our flow network, enhancing our performance in large motion regions. Extensive experiments demonstrate that our HDRFlow outperforms previous methods on standard benchmarks. To the best of our knowledge, HDRFlow is the first real-time HDR video reconstruction method for video sequences captured with alternating exposures, capable of processing 720p resolution inputs at 25ms. Project website: https: https://openimaginglab.github.io/HDRFlow/.

Lensless cameras offer significant advantages in size, weight, and cost compared to traditional lens-based systems. Without a focusing lens, lensless cameras rely on computational algorithms to recover the scenes from multiplexed measurements. However, current algorithms struggle with inaccurate forward imaging models and insufficient priors to reconstruct high-quality images. To overcome these limitations, we introduce a novel two-stage approach for consistent and photorealistic lensless image reconstruction. The first stage of our approach ensures data consistency by focusing on accurately reconstructing the low-frequency content with a spatially varying deconvolution method that adjusts to changes in the Point Spread Function (PSF) across the camera's field of view. The second stage enhances photorealism by incorporating a generative prior from pre-trained diffusion models. By conditioning on the low-frequency content retrieved in the first stage, the diffusion model effectively reconstructs the high-frequency details that are typically lost in the lensless imaging process, while also maintaining image fidelity. Our method achieves a superior balance between data fidelity and visual quality compared to existing methods, as demonstrated with two popular lensless systems, PhlatCam and DiffuserCam.

Image Signal Processors (ISPs) convert raw sensor signals into digital images, which significantly influence the image quality and the performance of downstream computer vision tasks. Designing ISP pipeline and tuning ISP parameters are two key steps for building an imaging and vision system. To find optimal ISP configurations, recent works use deep neural networks as a proxy to search for ISP parameters or ISP pipelines. However, these methods are primarily designed to maximize the image quality, which are sub-optimal in the performance of high-level computer vision tasks such as detection, recognition, and tracking. Moreover, after training, the learned ISP pipelines are mostly fixed at the inference time, whose performance degrades in dynamic scenes. To jointly optimize ISP structures and parameters, we propose AdaptiveISP, a task-driven and scene-adaptive ISP. One key observation is that for the majority of input images, only a few processing modules are needed to improve the performance of downstream recognition tasks, and only a few inputs require more processing. Based on this, AdaptiveISP utilizes deep reinforcement learning to automatically generate an optimal ISP pipeline and the associated ISP parameters to maximize the detection performance. Experimental results show that AdaptiveISP not only surpasses the prior state-of-the-art methods for object detection but also dynamically manages the trade-off between detection performance and computational cost, especially suitable for scenes with large dynamic range variations. Project website: https://openimaginglab.github.io/AdaptiveISP/.

Video Frame Interpolation (VFI) is important for video enhancement, frame rate up-conversion, and slow-motion generation. The introduction of event cameras, which capture per-pixel brightness changes asynchronously, has significantly enhanced VFI capabilities, particularly for high-speed, nonlinear motions. However, these event-based methods encounter challenges in low-light conditions, notably trailing artifacts and signal latency, which hinder their direct applicability and generalization. Addressing these issues, we propose a novel per-scene optimization strategy tailored for low-light conditions. This approach utilizes the internal statistics of a sequence to handle degraded event data under low-light conditions, improving the generalizability to different lighting and camera settings. To evaluate its robustness in low-light condition, we further introduce EVFI-LL, a unique RGB+Event dataset captured under low-light conditions. Our results demonstrate state-of-the-art performance in low-light environments. Project page: https://naturezhanghn.github.io/sim2real.

Due to the difficulty in collecting large-scale and perfectly aligned paired training data for Under-Display Camera (UDC) image restoration, previous methods resort to monitor-based image systems or simulation-based methods, sacrificing the realness of the data and introducing domain gaps. In this work, we revisit the classic stereo setup for training data collection – capturing two images of the same scene with one UDC and one standard camera. The key idea is to “copy” details from a high-quality reference image and “paste” them on the UDC image. While being able to generate real training pairs, this setting is susceptible to spatial misalignment due to perspective and depth of field changes. The problem is further compounded by the large domain discrepancy between the UDC and normal images, which is unique to UDC restoration. In this paper, we mitigate the non-trivial domain discrepancy and spatial misalignment through a novel Transformer-based framework that generates well-aligned yet high-quality target data for the corresponding UDC input. This is made possible through two carefully designed components, namely, the Domain Alignment Module (DAM) and Geometric Alignment Module (GAM), which encourage robust and accurate discovery of correspondence between the UDC and normal views. Extensive experiments show that high-quality and well-aligned pseudo UDC training pairs are beneficial for training a robust restoration network. Code and the dataset are available at https://github.com/jnjaby/AlignFormer.

Controllable image denoising aims to generate clean samples with human perceptual priors and balance sharpness and smoothness. In traditional filter-based denoising methods, this can be easily achieved by adjusting the filtering strength. However, for NN (Neural Network)-based models, adjusting the final denoising strength requires performing network inference each time, making it almost impossible for real-time user interaction. In this paper, we introduce Real-time Controllable Denoising (RCD), the first deep image and video denoising pipeline that provides a fully controllable user interface to edit arbitrary denoising levels in real-time with only one-time network inference. Unlike existing controllable denoising methods that require multiple denoisers and training stages, RCD replaces the last output layer (which usually outputs a single noise map) of an existing CNN-based model with a lightweight module that outputs multiple noise maps. We propose a novel Noise Decorrelation process to enforce the orthogonality of the noise feature maps, allowing arbitrary noise level control through noise map interpolation. This process is network-free and does not require network inference. Our experiments show that RCD can enable real-time editable image and video denoising for various existing heavy-weight models without sacrificing their original performance.

Recent work on discrete generative priors, in the form of codebooks, has shown exciting performance for image reconstruction and restoration, as the discrete prior space spanned by the codebooks increases the robustness against diverse image degradations. Nevertheless, these methods require separate training of codebooks for different image categories, which limits their use to specific image categories only (e.g. face, architecture, etc.), and fail to handle arbitrary natural images. In this paper, we propose AdaCode for learning image-adaptive codebooks for class-agnostic image restoration. Instead of learning a single codebook for each image category, we learn a set of basis codebooks. Given an input image, AdaCode learns a weight map with and computes a weighted combination of these basis codebooks for adaptive image restoration. Intuitively, AdaCode is a more flexible and expressive discrete generative prior than previous work. Experimental results demonstrate that AdaCode achieves state-of-the-art performance on image reconstruction and restoration tasks, including image super-resolution and inpainting. Codes are released at https://github.com/kechunl/AdaCode.

We show that pre-trained Generative Adversarial Networks (GANs) such as StyleGAN and BigGAN can be used as a latent bank to improve the performance of image super-resolution. While most existing perceptual-oriented approaches attempt to generate realistic outputs through learning with adversarial loss, our method, Generative LatEnt bANk (GLEAN), goes beyond existing practices by directly leveraging rich and diverse priors encapsulated in a pre-trained GAN. But unlike prevalent GAN inversion methods that require expensive image-specific optimization at runtime, our approach only needs a single forward pass for restoration. GLEAN can be easily incorporated in a simple encoder-bank-decoder architecture with multi-resolution skip connections. Employing priors from different generative models allows GLEAN to be applied to diverse categories (e.g., human faces, cats, buildings, and cars). We further present a lightweight version of GLEAN, named LightGLEAN, which retains only the critical components in GLEAN. Notably, LightGLEAN consists of only 21% of parameters and 35% of FLOPs while achieving comparable image quality. We extend our method to different tasks including image colorization and blind image restoration, and extensive experiments show that our proposed models perform favorably in comparison to existing methods. Codes and models are available at https://github.com/open-mmlab/mmediting.

The increased resolution of real-world videos presents a dilemma between efficiency and accuracy for deep Video Quality Assessment (VQA). On the one hand, keeping the original resolution will lead to unacceptable computational costs. On the other hand, existing practices, such as resizing or cropping, will change the quality of original videos due to difference in details or loss of contents, and are henceforth harmful to quality assessment. With obtained insight from the studies of spatial-temporal redundancy in the human visual system, visual quality around a neighbourhood has high probability to be similar, and this motivates us to investigate an effective quality-sensitive neighbourhood representative sampling scheme for VQA. In this work, we propose a unified scheme, spatial-temporal grid mini-cube sampling (St-GMS), and the resultant samples are namedfragments. In St-GMS, full-resolution videos are first divided into mini-cubes with predefined spatial-temporal grids, then the temporal-aligned quality representatives are sampled to compose the fragments that serve as inputs for VQA. In addition, we design the Fragment Attention Network (FANet), a network architecture tailored specifically for fragments. With fragments and FANet, the proposedFAST-VQAandFasterVQA(with an improved sampling scheme) achieves up to 1612× efficiency than the existing state-of-the-art, meanwhile achieving significantly better performance on all relevant VQA benchmarks.

Existing convolutional neural networks widely adopt spatial down-/up-sampling for multi-scale modeling. However, spatial up-sampling operators (e.g., interpolation, transposed convolution, and un-pooling) heavily depend on local pixel attention, incapably exploring the global dependency. In contrast, the Fourier domain is in accordance with the nature of global modeling according to the spectral convolution theorem. Unlike the spatial domain that easily performs up-sampling with the property of local similarity, up-sampling in the Fourier domain is more challenging as it does not follow such a local property. In this study, we propose a theoretically feasible Deep Fourier Up-Sampling (FourierUp) to solve these issues. We revisit the relationships between spatial and Fourier domains and reveal the transform rules on the features of different resolutions in the Fourier domain, which provide key insights for FourierUp's designs. FourierUp as a generic operator consists of three key components: 2D discrete Fourier transform, Fourier dimension increase rules, and 2D inverse Fourier transform, which can be directly integrated with existing networks. Extensive experiments across multiple computer vision tasks, including object detection, image segmentation, image de-raining, image dehazing, and guided image super-resolution, demonstrate the consistent performance gains obtained by introducing our FourierUp. Code will be publicly available.

Low-light image enhancement (LLIE) aims at improving the perception or interpretability of an image captured in an environment with poor illumination. Recent advances in this area are dominated by deep learning-based solutions, where many learning strategies, network structures, loss functions, training data, etc. have been employed. In this paper, we provide a comprehensive survey to cover various aspects ranging from algorithm taxonomy to unsolved open issues. To examine the generalization of existing methods, we propose a low-light image and video dataset, in which the images and videos are taken by different mobile phones' cameras under diverse illumination conditions. Besides, for the first time, we provide a unified online platform that covers many popular LLIE methods, of which the results can be produced through a user-friendly web interface. In addition to qualitative and quantitative evaluation of existing methods on publicly available and our proposed datasets, we also validate their performance in face detection in the dark. This survey together with the proposed dataset and online platform could serve as a reference source for future study and promote the development of this research field. The proposed platform and dataset as well as the collected methods, datasets, and evaluation metrics are publicly available and will be regularly updated. Project page: https://www.mmlab-ntu.com/project/lliv_survey/index.html.

While humans can effortlessly transform complex visual scenes into simple words and the other way around by leveraging their high-level understanding of the content, conventional or the more recent learned image compression codecs do not seem to utilize the semantic meanings of visual content to their full potential. Moreover, they focus mostly on rate-distortion and tend to underperform in perception quality especially in low bitrate regime, and often disregard the performance of downstream computer vision algorithms, which is a fast-growing consumer group of compressed images in addition to human viewers. In this paper, we (1) present a generic framework that can enable any image codec to leverage high-level semantics and (2) study the joint optimization of perception quality and distortion. Our idea is that given any codec, we utilize high-level semantics to augment the low-level visual features extracted by it and produce essentially a new, semantic-aware codec. We propose a three-phase training scheme that teaches semantic-aware codecs to leverage the power of semantic to jointly optimize rate-perception-distortion (R-PD) performance. As an additional benefit, semantic-aware codecs also boost the performance of downstream computer vision algorithms. To validate our claim, we perform extensive empirical evaluations and provide both quantitative and qualitative results.

We show that pre-trained Generative Adversarial Networks (GANs), e.g., StyleGAN, can be used as a latent bank to improve the restoration quality of large-factor image super-resolution (SR). While most existing SR approaches attempt to generate realistic textures through learning with adversarial loss, our method, Generative LatEnt bANk (GLEAN), goes beyond existing practices by directly leveraging rich and diverse priors encapsulated in a pre-trained GAN. But unlike prevalent GAN inversion methods that require expensive image-specific optimization at runtime, our approach only needs a single forward pass to generate the upscaled image. GLEAN can be easily incorporated in a simple encoder-bank-decoder architecture with multi-resolution skip connections. Switching the bank allows the method to deal with images from diverse categories, e.g., cat, building, human face, and car. Images upscaled by GLEAN show clear improvements in terms of fidelity and texture faithfulness in comparison to existing methods as shown in Fig. 1.

Development of Under-Display Camera (UDC) systems provides a true bezel-less and notch-free viewing experience on smartphones (and TV, laptops, tablets), while allowing images to be captured from the selfie camera embedded underneath. In a typical UDC system, the microstructure of the semi-transparent organic light-emitting diode (OLED) pixel array attenuates and diffracts the incident light on the camera, resulting in significant image quality degradation. Oftentimes, noise, flare, haze, and blur can be observed in UDC images. In this work, we aim to analyze and tackle the aforementioned degradation problems. We define a physics-based image formation model to better understand the degradation. In addition, we utilize one of the world’s first commodity UDC smartphone prototypes to measure the real-world Point Spread Function (PSF) of the UDC system, and provide a model-based data synthesis pipeline to generate realistically degraded images. We specially design a new domain knowledge-enabled Dynamic Skip Connection Network (DISCNet) to restore the UDC images. We demonstrate the effectiveness of our method through extensive experiments on both synthetic and real UDC data. Our physics-based image formation model and proposed DISCNet can provide foundations for further exploration in UDC image restoration, and even for general diffraction artifact removal in a broader sense.1

Image Signal Processor (ISP) is a crucial component in digital cameras that transforms sensor signals into images for us to perceive and understand. Existing ISP designs always adopt a fixed architecture, e.g., several sequential modules connected in a rigid order. Such a fixed ISP architecture may be suboptimal for real-world applications, where camera sensors, scenes and tasks are diverse. In this study, we propose a novel Reconfigurable ISP (ReconfigISP) whose architecture and parameters can be automatically tailored to specific data and tasks. In particular, we implement several ISP modules, and enable back-propagation for each module by training a differentiable proxy, hence allowing us to leverage the popular differentiable neural architecture search and effectively search for the optimal ISP architecture. A proxy tuning mechanism is adopted to maintain the accuracy of proxy networks in all cases. Extensive experiments conducted on image restoration and object detection, with different sensors, light conditions and efficiency constraints, validate the effectiveness of ReconfigISP. Only hundreds of parameters need tuning for every task1.

Image and video enhancement such as color constancy, low light enhancement, and tone mapping on smartphones is challenging, because high-quality images should be achieved efficiently with a limited resource budget. Unlike prior works that either used very deep CNNs or large Trans-former models, we propose a structure-aware lightweight Transformer, termed STAR, for real-time image enhancement. STAR is formulated to capture long-range dependencies between image patches, which naturally and implicitly captures the structural relationships of different regions in an image. STAR is a general architecture that can be easily adapted to different image enhancement tasks. Extensive experiments show that STAR can effectively boost the quality and efficiency of many tasks such as illumination enhancement, auto white balance, and photo retouching, which are indispensable components for image processing on smartphones. For example, STAR reduces model complexity and improves image quality compared to the recent state-of-the-art [19] on the MIT-Adobe FiveK dataset [7] (i.e., 1.8dB PSNR improvements with 25% parameters and 13% float operations.)

Model quantization is challenging due to many tedious hyper-parameters such as precision (bitwidth), dynamic range (minimum and maximum discrete values) and stepsize (interval between discrete values). Unlike prior arts that carefully tune these values, we present a fully differentiable approach to learn all of them, named Differentiable Dynamic Quantization (DDQ), which has several benefits. (1) DDQ is able to quantize challenging lightweight architectures like MobileNets, where different layers prefer different quantization parameters. (2) DDQ is hardware-friendly and can be easily implemented using low-precision matrix-vector multiplication, making it capable in many hardware such as ARM. (3) Extensive experiments show that DDQ outperforms prior arts on many networks and benchmarks, especially when models are already efficient and compact. e.g., DDQ is the first approach that achieves lossless 4-bit quantization for MobileNetV2 on ImageNet.

Time-of-Flight (ToF) cameras require active illumination to obtain depth information thus the power of illumination directly affects the performance of ToF cameras. Traditional ToF imaging algorithms are very sensitive to illumination and the depth accuracy degenerates rapidly with the power of it. Therefore, the design of a power efficient ToF camera always creates a painful dilemma for the illumination and the performance trade-off. In this paper, we show that despite the weak signals in many areas under extreme short exposure setting, these signals as a whole can be well utilized through a learning process which directly translates the weak and noisy ToF camera raw to depth map. This creates an opportunity to tackle the aforementioned dilemma and make a very power efficient ToF camera possible. To enable the learning, we collect a comprehensive dataset under a variety of scenes and photographic conditions by a specialized ToF camera. Experiments show that our method is able to robustly process ToF camera raw with the exposure time of one order of magnitude shorter than that used in conventional ToF cameras. In addition to evaluating our approach both quantitatively and qualitatively, we also discuss its implication to designing the next generation power efficient ToF cameras.

Multi-view Stereo (MVS) is a common solution in photogrammetry applications for the dense reconstruction of a static scene from images. The static scene assumption, however, limits the general applicability of MVS algorithms, as many day-to-day scenes undergo non-rigid motion, e.g., clothes, faces, or human bodies. In this paper, we open up a new challenging direction: Dense 3D reconstruction of scenes with non-rigid changes observed from a small number of images sparsely captured from different views with a single monocular camera, which we call non-rigid multi-view stereo (NRMVS) problem. We formulate this problem as a joint optimization of deformation and depth estimation, using deformation graphs as the underlying representation. We propose a new sparse 3D to 2D matching technique with a dense patch-match evaluation scheme to estimate the most plausible deformation field satisfying depth and photometric consistency. We show that a dense reconstruction of a scene with non-rigid changes from a few images is possible, and demonstrate that our method can be used to interpolate novel deformed scenes from various combinations of deformation estimates derived from the sparse views.

This paper proposes a deep neural architecture, PlaneRCNN, that detects and reconstructs piecewise planar regions from a single RGB image. PlaneRCNN employs a variant of Mask R-CNN to detect planes with their plane parameters and segmentation masks. PlaneRCNN then refines an arbitrary number of segmentation masks with a novel loss enforcing the consistency with a nearby view during training. The paper also presents a new benchmark with more fine-grained plane segmentations in the ground-truth, in which, PlaneRCNN outperforms existing state-of-the-art methods with significant margins in the plane detection, segmentation, and reconstruction metrics. PlaneRCNN makes an important step towards robust plane extraction method, which would have immediate impact on a wide range of applications including Robotics, Augmented Reality, and Virtual Reality.

Depth sensing is crucial for 3D reconstruction and scene understanding. Active depth sensors provide dense metric measurements, but often suffer from limitations such as restricted operating ranges, low spatial resolution, sensor interference, and high power consumption. In this paper, we propose a deep learning (DL) method to estimate per-pixel depth and its uncertainty continuously from a monocular video stream, with the goal of effectively turning an RGB camera into an RGB-D camera. Unlike prior DL-based methods, we estimate a depth probability distribution for each pixel rather than a single depth value, leading to an estimate of a 3D depth probability volume for each input frame. These depth probability volumes are accumulated over time under a Bayesian filtering framework as more incoming frames are processed sequentially, which effectively reduces depth uncertainty and improves accuracy, robustness, and temporal stability. Compared to prior work, the proposed approach achieves more accurate and stable results, and generalizes better to new datasets. Experimental results also show the output of our approach can be directly fed into classical RGB-D based 3D scanning methods for 3D scene reconstruction.

Inverse rendering aims to estimate physical attributes of a scene, e.g., reflectance, geometry, and lighting, from image(s). Inverse rendering has been studied primarily for single objects or with methods that solve for only one of the scene attributes. We propose the first learning based approach that jointly estimates albedo, normals, and lighting of an indoor scene from a single image. Our key contribution is the Residual Appearance Renderer (RAR), which can be trained to synthesize complex appearance effects (e.g., inter-reflection, cast shadows, near-field illumination, and realistic shading), which would be neglected otherwise. This enables us to perform self-supervised learning on real data using a reconstruction loss, based on re-synthesizing the input image from the estimated components. We finetune with real data after pretraining with synthetic data. To this end, we use physically-based rendering to create a large-scale synthetic dataset, named SUNCG-PBR, which is a significant improvement over prior datasets. Experimental results show that our approach outperforms state-of-the-art methods that estimate one or more scene attributes.

Maps are a key component in image-based camera localization and visual SLAM systems: they are used to establish geometric constraints between images, correct drift in relative pose estimation, and relocalize cameras after lost tracking. The exact definitions of maps, however, are often application-specific and hand-crafted for different scenarios (e.g. 3D landmarks, lines, planes, bags of visual words). We propose to represent maps as a deep neural net called MapNet, which enables learning a data-driven map representation. Unlike prior work on learning maps, MapNet exploits cheap and ubiquitous sensory inputs like visual odometry and GPS in addition to images and fuses them together for camera localization. Geometric constraints expressed by these inputs, which have traditionally been used in bundle adjustment or pose-graph optimization, are formulated as loss terms in MapNet training and also used during inference. In addition to directly improving localization accuracy, this allows us to update the MapNet (i.e., maps) in a self-supervised manner using additional unlabeled video sequences from the scene. We also propose a novel parameterization for camera rotation which is better suited for deep-learning based camera pose regression. Experimental results on both the indoor 7-Scenes dataset and the outdoor Oxford RobotCar dataset show significant performance improvement over prior work. The MapNet project webpage is https://goo.gl/mRB3Au.

Motion blur is a fundamental problem in computer vision as it impacts image quality and hinders inference. Traditional deblurring algorithms leverage the physics of the image formation model and use hand-crafted priors: they usually produce results that better reflect the underlying scene, but present artifacts. Recent learning-based methods implicitly extract the distribution of natural images directly from the data and use it to synthesize plausible images. Their results are impressive, but they are not always faithful to the content of the latent image. We present an approach that bridges the two. Our method fine-tunes existing deblurring neural networks in a self-supervised fashion by enforcing that the output, when blurred based on the optical flow between subsequent frames, matches the input blurry image. We show that our method significantly improves the performance of existing methods on several datasets both visually and in terms of image quality metrics.

Learning to insert an object instance into an image in a semantically coherent manner is a challenging and interesting problem. Solving it requires (a) determining a location to place an object in the scene and (b) determining its appearance at the location. Such an object insertion model can potentially facilitate numerous image editing and scene parsing applications. In this paper, we propose an end-to-end trainable neural network for the task of inserting an object instance mask of a specified class into the semantic label map of an image. Our network consists of two generative modules where one determines where the inserted object mask should be (i.e., location and scale) and the other determines what the object mask shape (and pose) should look like. The two modules are connected together via a spatial transformation network and jointly trained. We devise a learning procedure that leverage both supervised and unsupervised data and show our model can insert an object at diverse locations with various appearances. We conduct extensive experimental validations with comparisons to strong baselines to verify the effectiveness of the proposed network.

Multi-view stereo relies on feature correspondences for 3D reconstruction, and thus is fundamentally flawed in dealing with featureless scenes. In this paper, we propose polarimetric multi-view stereo, which combines per-pixel photometric information from polarization with epipolar constraints from multiple views for 3D reconstruction. Polarization reveals surface normal information, and is thus helpful to propagate depth to featureless regions. Polarimetric multi-view stereo is completely passive and can be applied outdoors in uncontrolled illumination, since the data capture can be done simply with either a polarizer or a polarization camera. Unlike previous work on shape-from-polarization which is limited to either diffuse polarization or specular polarization only, we propose a novel polarization imaging model that can handle real-world objects with mixed polarization. We prove there are exactly two types of ambiguities on estimating surface azimuth angles from polarization, and we resolve them with graph optimization and iso-depth contour tracing. This step significantly improves the initial depth map estimate, which are later fused together for complete 3D reconstruction. Extensive experimental results demonstrate high-quality 3D reconstruction and better performance than state-of-the-art multi-view stereo methods, especially on featureless 3D objects, such as ceramic tiles, office room with white walls, and highly reflective cars in the outdoors.

Facial analysis in videos, including head pose estimation and facial landmark localization, is key for many applications such as facial animation capture, human activity recognition, and human-computer interaction. In this paper, we propose to use a recurrent neural network (RNN) for joint estimation and tracking of facial features in videos. We are inspired by the fact that the computation performed in an RNN bears resemblance to Bayesian filters, which have been used for tracking in many previous methods for facial analysis from videos. Bayesian filters used in these methods, however, require complicated, problem-specific design and tuning. In contrast, our proposed RNN-based method avoids such tracker-engineering by learning from training data, similar to how a convolutional neural network (CNN) avoids feature-engineering for image classification. As an end-to-end network, the proposed RNN-based method provides a generic and holistic solution for joint estimation and tracking of various types of facial features from consecutive video frames. Extensive experimental results on head pose estimation and facial landmark localization from videos demonstrate that the proposed RNN-based method outperforms frame-wise models and Bayesian filtering. In addition, we create a large-scale synthetic dataset for head pose estimation, with which we achieve state-of-the-art performance on a benchmark dataset.

Estimating surface reflectance (BRDF) is one key component for complete 3D scene capture, with wide applications in virtual reality, augmented reality, and human computer interaction. Prior work is either limited to controlled environments (e.g., gonioreflectometers, light stages, or multi-camera domes), or requires the joint optimization of shape, illumination, and reflectance, which is often computationally too expensive (e.g., hours of running time) for real-time applications. Moreover, most prior work requires HDR images as input which further complicates the capture process. In this paper, we propose a lightweight approach for surface reflectance estimation directly from 8-bit RGB images in real-time, which can be easily plugged into any 3D scanning-and-fusion system with a commodity RGBD sensor. Our method is learning-based, with an inference time of less than 90ms per scene and a model size of less than 340K bytes. We propose two novel network architectures, HemiCNN and Grouplet, to deal with the unstructured input data from multiple viewpoints under unknown illumination. We further design a loss function to resolve the color-constancy and scale ambiguity. In addition, we have created a large synthetic dataset, SynBRDF, which comprises a total of 500K RGBD images rendered with a physically-based ray tracer under a variety of natural illumination, covering 5000 materials and 5000 shapes. SynBRDF is the first large-scale benchmark dataset for reflectance estimation. Experiments on both synthetic data and real data show that the proposed method effectively recovers surface reflectance, and outperforms prior work for reflectance estimation in uncontrolled environments.

In this paper, we propose a spatial propagation networks for learning affinity matrix. We show that by constructing a row/column linear propagation model, the spatially variant transformation matrix constitutes an affinity matrix that models dense, global pairwise similarities of an image. Specifically, we develop a three-way connection for the linear propagation model, which (a) formulates a sparse transformation matrix where all elements can be the output from a deep CNN, but (b) results in a dense affinity matrix that is effective to model any task-specific pairwise similarity. Instead of designing the similarity kernels according to image features of two points, we can directly output all similarities in a pure data-driven manner. The spatial propagation network is a generic framework that can be applied to numerous tasks, which traditionally benefit from designed affinity, e.g., image matting, colorization, and guided filtering, to name a few. Furthermore, the model can also learn semantic-aware affinity for high-level vision tasks due to the learning capability of the deep model. We validate the proposed framework by refinement of object segmentation. Experiments on the HELEN face parsing and PASCAL VOC-2012 semantic segmentation tasks show that the spatial propagation network provides general, effective and efficient solutions for generating high-quality segmentation results.

Classifying raw, unpainted materials--metal, plastic, ceramic, fabric, and so on--is an important yet challenging task for computer vision. Previous works measure subsets of surface spectral reflectance as features for classification. However, acquiring the full spectral reflectance is time consuming and error-prone. In this paper, we propose to use coded illumination to directly measure discriminative features for material classification. Optimal illumination patterns--which we call "discriminative illumination"--are learned from training samples, after projecting to which the spectral reflectance of different materials are maximally separated. This projection is automatically realized by the integration of incident light for surface reflection. While a single discriminative illumination is capable of linear, two-class classification, we show that multiple discriminative illuminations can be used for nonlinear and multiclass classification. We also show theoretically that the proposed method has higher signal-to-noise ratio than previous methods due to light multiplexing. Finally, we construct an LED-based multispectral dome and use the discriminative illumination method for classifying a variety of raw materials, including metal (aluminum, alloy, steel, stainless steel, brass, and copper), plastic, ceramic, fabric, and wood. Experimental results demonstrate its effectiveness.

Cameras face a fundamental trade-off between spatial and temporal resolution. Digital still cameras can capture images with high spatial resolution, but most high-speed video cameras have relatively low spatial resolution. It is hard to overcome this trade-off without incurring a significant increase in hardware costs. In this paper, we propose techniques for sampling, representing, and reconstructing the space-time volume to overcome this trade-off. Our approach has two important distinctions compared to previous works: 1) We achieve sparse representation of videos by learning an overcomplete dictionary on video patches, and 2) we adhere to practical hardware constraints on sampling schemes imposed by architectures of current image sensors, which means that our sampling function can be implemented on CMOS image sensors with modified control units in the future. We evaluate components of our approach, sampling function and sparse representation, by comparing them to several existing approaches. We also implement a prototype imaging system with pixel-wise coded exposure control using a liquid crystal on silicon device. System characteristics such as field of view and modulation transfer function are evaluated for our imaging system. Both simulations and experiments on a wide range of scenes show that our method can effectively reconstruct a video from a single coded image while maintaining high spatial resolution.

We present a computational imaging method for raw material classification using features of Bidirectional Texture Functions (BTF). Texture is an intrinsic feature for many materials, such as wood, fabric, and granite. At appropriate scales, even "uniform" materials will also exhibit texture features that can be helpful for recognition, such as paper, metal, and ceramic. To cope with the high-dimensionality of BTFs, in this paper, we proposed to learn discriminative illumination patterns and texture filters, with which we can directly measure optimal projections of BTFs for classification. We also studied the effects of texture rotation and scale variation for material classification. We built an LED-based multispectral dome, with which we have acquired a BTF database of a variety of materials and demonstrated the effectiveness of the proposed approach for material classification.

Camera spectral sensitivity functions relate scene radiance with captured RGB triplets. They are important for many computer vision tasks that use color information, such as multispectral imaging, color rendering, and color constancy. In this paper, we aim to explore the space of spectral sensitivity functions for digital color cameras. After collecting a database of 28 cameras covering a variety of types, we find this space convex and two-dimensional. Based on this statistical model, we propose two methods to recover camera spectral sensitivities using regular reflective color targets (e.g., color checker) from a single image with and without knowing the illumination. We show the proposed model is more accurate and robust for estimating camera spectral sensitivities than other basis functions. We also show two applications for the recovery of camera spectral sensitivities - simulation of color rendering for cameras and computational color constancy.

We propose a new method named compressive structured light for recovering inhomogeneous participating media. Whereas conventional structured light methods emit coded light patterns onto the surface of an opaque object to establish correspondence for triangulation, compressive structured light projects patterns into a volume of participating medium to produce images which are integral measurements of the volume density along the line of sight. For a typical participating medium encountered in the real world, the integral nature of the acquired images enables the use of compressive sensing techniques that can recover the entire volume density from only a few measurements. This makes the acquisition process more efficient and enables reconstruction of dynamic volumetric phenomena. Moreover, our method requires the projection of multiplexed coded illumination, which has the added advantage of increasing the signal-to-noise ratio of the acquisition. Finally, we propose an iterative algorithm to correct for the attenuation of the participating medium during the reconstruction process. We show the effectiveness of our method with simulations as well as experiments on the volumetric recovery of multiple translucent layers, 3D point clouds etched in glass, and the dynamic process of milk drops dissolving in water.

Classifying raw, unpainted materials - metal, plastic, ceramic, fabric, etc. - is an important yet challenging task for computer vision. Previous works measure subsets of surface spectral reflectance as features for classification. However, acquiring the full spectral reflectance is time-consuming and error-prone. In this paper, we propose to use coded illumination to directly measure discriminative features for material classification. Optimal illumination patterns - which we call “discriminative illumination” - are learned from training samples, after projecting to which, the spectral reflectance of different materials are maximally separated. This projection is automatically realized by the integration of incident light for surface reflection. While a single discriminative illumination is capable of linear, two-class classification, we show that multiple discriminative illuminations can be used for nonlinear and multi-class classification. We also show theoretically the proposed method has higher signal-to-noise ratio than previous methods due to light multiplexing. Finally, we construct a LED-based multi-spectral dome and use the discriminative illumination method for classifying a variety of raw materials, including metal (aluminum, alloy, steel, stainless steel, brass and copper), plastic, ceramic, fabric and wood. Experimental results demonstrate the effectiveness of the proposed method.

Global illumination effects such as inter-reflections and subsurface scattering result in systematic, and often significant errors in scene recovery using active illumination. Recently, it was shown that the direct and global components could be separated efficiently for a scene illuminated with a single light source. In this paper, we study the problem of direct-global separation for multiple light sources. We derive a theoretical lower bound for the number of required images, and propose a multiplexed illumination scheme which achieves this lower bound. We analyze the signal-to-noise ratio (SNR) characteristics of the proposed illumination multiplexing method in the context of direct-global separation. We apply our method to several scene recovery techniques requiring multiple light sources, including shape from shading, structured light 3D scanning, photometric stereo, and reflectance estimation. Both simulation and experimental results show that the proposed method can accurately recover scene information with fewer images compared to sequentially separating direct-global components for each light source.

Cameras face a fundamental tradeoff between the spatial and temporal resolution - digital still cameras can capture images with high spatial resolution, but most high-speed video cameras suffer from low spatial resolution. It is hard to overcome this tradeoff without incurring a significant increase in hardware costs. In this paper, we propose techniques for sampling, representing and reconstructing the space-time volume in order to overcome this tradeoff. Our approach has two important distinctions compared to previous works: (1) we achieve sparse representation of videos by learning an over-complete dictionary on video patches, and (2) we adhere to practical constraints on sampling scheme which is imposed by architectures of present image sensor devices. Consequently, our sampling scheme can be implemented on image sensors by making a straightforward modification to the control unit. To demonstrate the power of our approach, we have implemented a prototype imaging system with per-pixel coded exposure control using a liquid crystal on silicon (LCoS) device. Using both simulations and experiments on a wide range of scenes, we show that our method can effectively reconstruct a video from a single image maintaining high spatial resolution.

Fingerprint analysis is typically based on the location and pattern of detected singular points in the images. These singular points (cores and deltas) not only represent the characteristics of local ridge patterns but also determine the topological structure (i.e., fingerprint type) and largely influence the orientation field. In this paper, we propose a novel algorithm for singular points detection. After an initial detection using the conventional Poincaré Index method, a so-called DORIC feature is used to remove spurious singular points. Then, the optimal combination of singular points is selected to minimize the difference between the original orientation field and the model-based orientation field reconstructed using the singular points. A core-delta relation is used as a global constraint for the final selection of singular points. Experimental results show that our algorithm is accurate and robust, giving better results than competing approaches. The proposed detection algorithm can also be used for more general 2D oriented patterns, such as fluid flow motion, and so forth.

Dirt on camera lenses, and occlusions from thin objects such as fences, are two important types of artifacts in digital imaging systems. These artifacts are not only an annoyance for photographers, but also a hindrance to computer vision and digital forensics. In this paper, we show that both effects can be described by a single image formation model, wherein an intermediate layer (of dust, dirt or thin occluders) both attenuates the incoming light and scatters stray light towards the camera. Because of camera defocus, these artifacts are low-frequency and either additive or multiplicative, which gives us the power to recover the original scene radiance pointwise. We develop a number of physics-based methods to remove these effects from digital photographs and videos. For dirty camera lenses, we propose two methods to estimate the attenuation and the scattering of the lens dirt and remove the artifacts -- either by taking several pictures of a structured calibration pattern beforehand, or by leveraging natural image statistics for post-processing existing images. For artifacts from thin occluders, we propose a simple yet effective iterative method that recovers the original scene from multiple apertures. The method requires two images if the depths of the scene and the occluder layer are known, or three images if the depths are unknown. The effectiveness of our proposed methods are demonstrated by both simulated and real experimental results.

Although real-world surfaces can exhibit significant variation in materials --- glossy, diffuse, metallic, etc. --- printers are usually used to reproduce color or gray-scale images. We propose a complete system that uses appropriate inks and foils to print documents with a variety of material properties. Given a set of inks with known Bidirectional Reflectance Distribution Functions (BRDFs), our system automatically finds the optimal linear combinations to approximate the BRDFs of the target documents. Novel gamut-mapping algorithms preserve the relative glossiness between different BRDFs, and halftoning is used to produce patterns to be sent to the printer. We demonstrate the effectiveness of this approach with printed samples of a number of measured spatially-varying BRDFs.

As an important feature, orientation field describes the global structure of fingerprints. It provides robust discriminatory information other than traditional widely-used minutiae points. However, there are few works explicitly incorporating this information into fingerprint matching stage, partly due to the difficulty of saving the orientation field in the feature template. In this paper, we propose a novel representation for fingerprints which includes both minutiae and model-based orientation field. Then, fingerprint matching can be done by combining the decisions of the matchers based on the global structure (orientation field) and the local cue (minutiae). We have conducted a set of experiments on large-scale databases and made thorough comparisons with the state-of-the-arts. Extensive experimental results show that combining these local and global discriminative information can largely improve the performance. The proposed system is more robust and accurate than conventional minutiae-based methods, and also better than the previous works which implicitly incorporate the orientation information. In this system, the feature template takes less than 420 bytes, and the feature extraction and matching procedures can be done in about 0.30 s. We also show that the global orientation field is beneficial to the alignment of the fingerprints which are either incomplete or poor-qualitied.

Content-based image retrieval (CBIR) has been more and more important in the last decade, and the gap between high-level semantic concepts and low-level visual features hinders further performance improvement. The problem of online feature selection is critical to really bridge this gap. In this paper, we investigate online feature selection in the relevance feedback learning process to improve the retrieval performance of the region-based image retrieval system. Our contributions are mainly in three areas. 1) A novel feature selection criterion is proposed, which is based on the psychological similarity between the positive and negative training sets. 2) An effective online feature selection algorithm is implemented in a boosting manner to select the most representative features for the current query concept and combine classifiers constructed over the selected features to retrieve images. 3) To apply the proposed feature selection method in region-based image retrieval systems, we propose a novel region-based representation to describe images in a uniform feature space with real-valued fuzzy features. Our system is suitable for online relevance feedback learning in CBIR by meeting the three requirements: learning with small size training set, the intrinsic asymmetry property of training samples, and the fast response requirement. Extensive experiments, including comparisons with many state-of-the-arts, show the effectiveness of our algorithm in improving the retrieval performance and saving the processing time.

For computer graphics rendering, we generally assume that the appearance of surfaces remains static over time. Yet, there are a number of natural processes that cause surface appearance to vary dramatically, such as burning of wood, wetting and drying of rock and fabric, decay of fruit skins, and corrosion and rusting of steel and copper. In this paper, we take a significant step towards measuring, modeling, and rendering time-varying surface appearance. We describe the acquisition of the first time-varying database of 26 samples, encompassing a variety of natural processes including burning, drying, decay, and corrosion. Our main technical contribution is a Space-Time Appearance Factorization (STAF). This model factors space and time-varying effects. We derive an overall temporal appearance variation characteristic curve of the specific process, as well as space-dependent textures, rates, and offsets. This overall temporal curve controls different spatial locations evolve at the different rates, causing spatial patterns on the surface over time. We show that the model accurately represents a variety of phenomena. Moreover, it enables a number of novel rendering applications, such as transfer of the time-varying effect to a new static surface, control to accelerate time evolution in certain areas, extrapolation beyond the acquired sequence, and texture synthesis of time-varying appearance.

To really bridge the gap between high-level semantics and low-level features in content-based image retrieval (CBIR), a problem that must be solved is which features are suitable for explaining the current query concept. We propose a novel feature selection (FS) criterion based on a psychological similarity measurement, generalized feature contrast model, and implement an online feature selection algorithm in a boosting manner to select the most representative features and do classification during each feedback round. The advantages of the proposed method are: it does not require a Gaussian assumption for "relevant" images as other online FS methods do; it accounts for the intrinsic asymmetry between "relevant" and "irrelevant" image sets in CBIR online learning; it is very fast. Extensive experiments have shown our algorithm's effectiveness.

As a global feature of fingerprints, the orientation field is very important for automatic fingerprint recognition. Many algorithms have been proposed for orientation field estimation, but their results are unsatisfactory, especially for poor quality fingerprint images. In this paper, a model-based method for the computation of orientation field is proposed. First a combination model is established for the representation of the orientation field by conidering its smoothness except for several singular points, in which a polynomial model is used to describe the orientation field globally and a point-charge model is taken to improve the accuracy locally at each singular point. When the coarse field is computed by using the gradient-based algorithm, a further result can be gained by using the model for a weighted approximation. Due to the global approximation, this model-based orientation field estimation algorithm has a robust performance on different fingerprint images. A further experiment shows that the performance of a whole fingerprint recognition system can be improved by applying this algorithm instead of previous orientation estimation methods.

As a global feature of the fingerprint, the orientation field is very important to an automatic fingerprint identification system (AFIS). Establishing an accurate and precise model for the orientation field will not only improve the performance of orientation estimation, but will also make it feasible to apply orientation information into the matching process. Such a novel model for the orientation field for a fingerprint is proposed. We use a polynomial model to approximate the orientation field globally and a point-charge model at each singular point to improve the approximation locally. These two models are combined together by a weight function. Experimental results are provided to illustrate that this combination model is more accurate and robust to noise compared with previous works. Its applications are also discussed.

Orientation field estimation is very important to automatic fingerprint recognition system. In this paper, a model-based algorithm for orientation field estimation is proposed. First the coarse field is computed by using conventional algorithms. Considering that the real orientation field is rather smooth except for several singular points, a combination model is proposed to approximate the coarse field. Experimental results are provided to illustrate the final estimation result is rather accurate and robust.