arXiv subjects

Soo-Chang Pei

Publications and source records attributed to Soo-Chang Pei.

16 recordsLinked to original sources

arXiv2025

BARIS: Boundary-Aware Refinement with Environmental Degradation Priors for Robust Underwater Instance Segmentation

Underwater instance segmentation is challenging due to adverse visual conditions such as light attenuation, scattering, and color distortion, which degrade model performance. In this work, we propose BARIS-Decoder (Boundary-Aware Refinement Decoder for Instance Segmentation), a framework that enhances segmentation accuracy through feature refinement. To address underwater degradations, we introduce the Environmental Robust Adapter (ERA), which efficiently models underwater degradation patterns while reducing trainable parameters by over 90\% compared to full fine-tuning. The integration of BARIS-Decoder with ERA-tuning, referred to as BARIS-ERA, achieves state-of-the-art performance, surpassing Mask R-CNN by 3.4 mAP with a Swin-B backbone and 3.8 mAP with ConvNeXt V2. Our findings demonstrate the effectiveness of BARIS-ERA in advancing underwater instance segmentation, providing a robust and efficient solution.

cs.CV↗

arXiv2024

Always Clear Days: Degradation Type and Severity Aware All-In-One Adverse Weather Removal

All-in-one adverse weather removal is an emerging topic on image restoration, which aims to restore multiple weather degradations in an unified model, and the challenge are twofold. First, discover and handle the property of multi-domain in target distribution formed by multiple weather conditions. Second, design efficient and effective operations for different degradations. To resolve this problem, most prior works focus on the multi-domain caused by different weather types. Inspired by inter\&intra-domain adaptation literature, we observe that not only weather type but also weather severity introduce multi-domain within each weather type domain, which is ignored by previous methods, and further limit their performance. To this end, we propose a degradation type and severity aware model, called UtilityIR, for blind all-in-one bad weather image restoration. To extract weather information from single image, we propose a novel Marginal Quality Ranking Loss (MQRL) and utilize Contrastive Loss (CL) to guide weather severity and type extraction, and leverage a bag of novel techniques such as Multi-Head Cross Attention (MHCA) and Local-Global Adaptive Instance Normalization (LG-AdaIN) to efficiently restore spatial varying weather degradation. The proposed method can outperform the state-of-the-art methods subjectively and objectively on different weather removal tasks with a large margin, and enjoy less model parameters. Proposed method even can restore unseen combined multiple degradation images, and modulate restoration level. Implementation code and pre-trained weights will be available at \url{https://github.com/fordevoted/UtilityIR}

cs.CV↗

arXiv2023

Panoptic-Depth Color Map for Combination of Depth and Image Segmentation

Image segmentation and depth estimation are crucial tasks in computer vision, especially in autonomous driving scenarios. Although these tasks are typically addressed separately, we propose an innovative approach to combine them in our novel deep learning network, Panoptic-DepthLab. By incorporating an additional depth estimation branch into the segmentation network, it can predict the depth of each instance segment. Evaluating on Cityscape dataset, we demonstrate the effectiveness of our method in achieving high-quality segmentation results with depth and visualize it with a color map. Our proposed method demonstrates a new possibility of combining different tasks and networks to generate a more comprehensive image recognition result to facilitate the safety of autonomous driving vehicles.

cs.CV↗

arXiv2023

Perspective-aware Convolution for Monocular 3D Object Detection

Monocular 3D object detection is a crucial and challenging task for autonomous driving vehicle, while it uses only a single camera image to infer 3D objects in the scene. To address the difficulty of predicting depth using only pictorial clue, we propose a novel perspective-aware convolutional layer that captures long-range dependencies in images. By enforcing convolutional kernels to extract features along the depth axis of every image pixel, we incorporates perspective information into network architecture. We integrate our perspective-aware convolutional layer into a 3D object detector and demonstrate improved performance on the KITTI3D dataset, achieving a 23.9\% average precision in the easy benchmark. These results underscore the importance of modeling scene clues for accurate depth inference and highlight the benefits of incorporating scene structure in network design. Our perspective-aware convolutional layer has the potential to enhance object detection accuracy by providing more precise and context-aware feature extraction.

cs.CV↗

arXiv2022

Domain Adaptation for Underwater Image Enhancement via Content and Style Separation

Underwater image suffer from color cast, low contrast and hazy effect due to light absorption, refraction and scattering, which degraded the high-level application, e.g, object detection and object tracking. Recent learning-based methods demonstrate astonishing performance on underwater image enhancement, however, most of these works use synthetic pair data for supervised learning and ignore the domain gap to real-world data. To solve this problem, we propose a domain adaptation framework for underwater image enhancement via content and style separation, different from prior works of domain adaptation for underwater image enhancement, which target to minimize the latent discrepancy of synthesis and real-world data, we aim to separate encoded feature into content and style latent and distinguish style latent from different domains, i.e. synthesis, real-world underwater and clean domain, and process domain adaptation and image enhancement in latent space. By latent manipulation, our model provide a user interact interface to adjust different enhanced level for continuous change. Experiment on various public real-world underwater benchmarks demonstrate that the proposed framework is capable to perform domain adaptation for underwater image enhancement and outperform various state-of-the-art underwater image enhancement algorithms in quantity and quality. The model and source code will be available at https://github.com/fordevoted/UIESS

cs.CV↗

arXiv2020

Explorable Tone Mapping Operators

Tone-mapping plays an essential role in high dynamic range (HDR) imaging. It aims to preserve visual information of HDR images in a medium with a limited dynamic range. Although many works have been proposed to provide tone-mapped results from HDR images, most of them can only perform tone-mapping in a single pre-designed way. However, the subjectivity of tone-mapping quality varies from person to person, and the preference of tone-mapping style also differs from application to application. In this paper, a learning-based multimodal tone-mapping method is proposed, which not only achieves excellent visual quality but also explores the style diversity. Based on the framework of BicycleGAN, the proposed method can provide a variety of expert-level tone-mapped results by manipulating different latent codes. Finally, we show that the proposed method performs favorably against state-of-the-art tone-mapping algorithms both quantitatively and qualitatively.

cs.CV↗

arXiv2017

Fast and accurate computation of normalized Bargmann transform

The linear canonical transform (LCT) was extended to complex-valued parameters, called complex LCT, to describe the complex amplitude propagation through lossy or lossless optical systems. Bargmann transform is a special case of the complex LCT. In this paper, we normalize the Bargmann transform such that it can be bounded near infinity. We derive the relationships of the normalized Bargmann transform to Gabor transform, Hermite-Gaussian functions, gyrator transform, and 2D nonseparable LCT. Several kinds of fast and accurate computational methods of the normalized Bargmann transform and its inverse are proposed based on these relationships.

quant-ph↗

arXiv2017

Reversible Joint Hilbert and Linear Canonical Transform Without Distortion

Generalized analytic signal associated with the linear canonical transform (LCT) was proposed recently by Fu and Li ["Generalized Analytic Signal Associated With Linear Canonical Transform," Opt. Commun., vol. 281, pp. 1468-1472, 2008]. However, most real signals, especially for baseband real signals, cannot be perfectly recovered from their generalized analytic signals. Therefore, in this paper, the conventional Hilbert transform (HT) and analytic signal associated with the LCT are concerned. To transform a real signal into the LCT of its HT, two integral transforms (i.e., the HT and LCT) are required. The goal of this paper is to simplify cascades of multiple integral transforms, which may be the HT, analytic signal, LCT or inverse LCT. The proposed transforms can reduce the complexity when realizing the relationships among the following six kinds of signals: a real signal, its HT and analytic signal, and the LCT of these three signals. Most importantly, all the proposed transforms are reversible and undistorted. Using the proposed transforms, several signal processing applications are discussed and show the advantages and flexibility over simply using the analytic signal or the LCT.

cs.IT↗

arXiv2017

Fast Discrete Linear Canonical Transform Based on CM-CC-CM Decomposition and FFT

In this paper, a discrete LCT (DLCT) irrelevant to the sampling periods and without oversampling operation is developed. This DLCT is based on the well-known CM-CC-CM decomposition, that is, implemented by two discrete chirp multiplications (CMs) and one discrete chirp convolution (CC). This decomposition doesn't use any scaling operation which will change the sampling period or cause the interpolation error. Compared with previous works, DLCT calculated by direct summation and DLCT based on center discrete dilated Hermite functions (CDDHFs), the proposed method implemented by FFTs has much lower computational complexity. The relation between the proposed DLCT and the continuous LCT is also derived to approximate the samples of the continuous LCT. Simulation results show that the proposed method somewhat outperforms the CDDHFs-based method in the approximation accuracy. Besides, the proposed method has approximate additivity property with error as small as the CDDHFs-based method. Most importantly, the proposed method has perfect reversibility, which doesn't hold in many existing DLCTs. With this property, it is unnecessary to develop the inverse DLCT additionally because it can be replaced by the forward DLCT.

cs.IT↗

arXiv2017

Discrete Gyrator Transforms: Computational Algorithms and Applications

As an extension of the 2D fractional Fourier transform (FRFT) and a special case of the 2D linear canonical transform (LCT), the gyrator transform was introduced to produce rotations in twisted space/spatial-frequency planes. It is a useful tool in optics, signal processing and image processing. In this paper, we develop discrete gyrator transforms (DGTs) based on the 2D LCT. Taking the advantage of the additivity property of the 2D LCT, we propose three kinds of DGTs, each of which is a cascade of low-complexity operators. These DGTs have different constraints, characteristics, and properties, and are realized by different computational algorithms. Besides, we propose a kind of DGT based on the eigenfunctions of the gyrator transform. This DGT is an orthonormal transform, and thus its comprehensive properties, especially the additivity property, make it more useful in many applications. We also develop an efficient computational algorithm to significantly reduce the complexity of this DGT. At the end, a brief review of some important applications of the DGTs is presented, including mode conversion, sampling and reconstruction, watermarking, and image encryption.

cs.CV↗

arXiv2017

Two-dimensional nonseparable discrete linear canonical transform based on CM-CC-CM-CC decomposition

As a generalization of the two-dimensional Fourier transform (2D FT) and 2D fractional Fourier transform, the 2D nonseparable linear canonical transform (2D NsLCT) is useful in optics, signal and image processing. To reduce the digital implementation complexity of the 2D NsLCT, some previous works decomposed the 2D NsLCT into several low-complexity operations, including 2D FT, 2D chirp multiplication (2D CM) and 2D affine transformations. However, 2D affine transformations will introduce interpolation error. In this paper, we propose a new decomposition called CM-CC-CM-CC decomposition, which decomposes the 2D NsLCT into two 2D CMs and two 2D chirp convolutions (2D CCs). No 2D affine transforms are involved. Simulation results show that the proposed methods have higher accuracy, lower computational complexity and smaller error in the additivity property compared with the previous works. Plus, the proposed methods have perfect reversibility property that one can reconstruct the input signal/image losslessly from the output.

cs.CV↗

arXiv2017

STFT with Adaptive Window Width Based on the Chirp Rate

An adaptive time-frequency representation (TFR) with higher energy concentration usually requires higher complexity. Recently, a low-complexity adaptive short-time Fourier transform (ASTFT) based on the chirp rate has been proposed. To enhance the performance, this method is substantially modified in this paper: i) because the wavelet transform used for instantaneous frequency (IF) estimation is not signal-dependent, a low-complexity ASTFT based on a novel concentration measure is addressed; ii) in order to increase robustness to IF estimation error, the principal component analysis (PCA) replaces the difference operator for calculating the chirp rate; and iii) a more robust Gaussian kernel with time-frequency-varying window width is proposed. Simulation results show that our method has higher energy concentration than the other ASTFTs, especially for multicomponent signals and nonlinear FM signals. Also, for IF estimation, our method is superior to many other adaptive TFRs in low signal-to-noise ratio (SNR) environments.

cs.IT↗

arXiv2016

Distributed Compressive Sensing: Performance Analysis with Diverse Signal Ensembles

Distributed compressive sensing is a framework considering jointly sparsity within signal ensembles along with multiple measurement vectors (MMVs). The current theoretical bound of performance for MMVs, however, is derived to be the same with that for single MV (SMV) because no assumption about signal ensembles is made. In this work, we propose a new concept of inducing the factor called "Euclidean distances between signals" for the performance analysis of MMVs. The novelty is that the size of signal ensembles will be taken into consideration in our analysis to theoretically show that MMVs indeed exhibit better performance than SMV. Although our concept can be broadly applied to CS algorithms with MMVs, the case study conducted on a well-known greedy solver called simultaneous orthogonal matching pursuit (SOMP) will be explored in this paper. We show that the performance of SOMP, when incorporated with our concept by modifying the steps of support detection and signal estimations, will be improved remarkably, especially when the Euclidean distances between signals are short. The performance of modified SOMP is verified to meet our theoretical prediction.

cs.IT↗

arXiv2016

Fast Greedy Approaches for Compressive Sensing of Large-Scale Signals

Cost-efficient compressive sensing is challenging when facing large-scale data, {\em i.e.}, data with large sizes. Conventional compressive sensing methods for large-scale data will suffer from low computational efficiency and massive memory storage. In this paper, we revisit well-known solvers called greedy algorithms, including Orthogonal Matching Pursuit (OMP), Subspace Pursuit (SP), Orthogonal Matching Pursuit with Replacement (OMPR). Generally, these approaches are conducted by iteratively executing two main steps: 1) support detection and 2) solving least square problem. To reduce the cost of Step 1, it is not hard to employ the sensing matrix that can be implemented by operator-based strategy instead of matrix-based one and can be speeded by fast Fourier Transform (FFT). Step 2, however, requires maintaining and calculating a pseudo-inverse of a sub-matrix, which is random and not structural, and, thus, operator-based matrix does not work. To overcome this difficulty, instead of solving Step 2 by a closed-form solution, we propose a fast and cost-effective least square solver, which combines a Conjugate Gradient (CG) method with our proposed weighted least square problem to iteratively approximate the ground truth yielded by a greedy algorithm. Extensive simulations and theoretical analysis validate that the proposed method is cost-efficient and is readily incorporated with the existing greedy algorithms to remarkably improve the performance for large-scale problems.

cs.DS↗

arXiv2016

Fast Binary Embedding via Circulant Downsampled Matrix -- A Data-Independent Approach

Binary embedding of high-dimensional data aims to produce low-dimensional binary codes while preserving discriminative power. State-of-the-art methods often suffer from high computation and storage costs. We present a simple and fast embedding scheme by first downsampling N-dimensional data into M-dimensional data and then multiplying the data with an MxM circulant matrix. Our method requires O(N +M log M) computation and O(N) storage costs. We prove if data have sparsity, our scheme can achieve similarity-preserving well. Experiments further demonstrate that though our method is cost-effective and fast, it still achieves comparable performance in image applications.

cs.IT↗

arXiv2015

Sparse Fast Fourier Transform for Exactly and Generally K-Sparse Signals by Downsampling and Sparse Recovery

Fast Fourier Transform (FFT) is one of the most important tools in digital signal processing. FFT costs O(N \log N) for transforming a signal of length N. Recently, Sparse Fourier Transform (SFT) has emerged as a critical issue addressing how to compute a compressed Fourier transform of a signal with complexity being related to the sparsity of its spectrum. In this paper, a new SFT algorithm is proposed for both exactly K-sparse signals (with K non-zero frequencies) and generally K-sparse signals (with K significant frequencies), with the assumption that the distribution of the non-zero frequencies is uniform. The nuclear idea is to downsample the input signal at the beginning; then, subsequent processing operates under downsampled signals, where signal lengths are proportional to O(K). Downsampling, however, possibly leads to "aliasing." By the shift property of DFT, we recast the aliasing problem as complex Bose-Chaudhuri-Hocquenghem (BCH) codes solved by syndrome decoding. The proposed SFT algorithm for exactly K-sparse signals recovers 1-τfrequencies with computational complexity O(K \log K) and probability at least 1-O(\frac{c}τ)^{τK} under K=O(N), where c is a user-controlled parameter. For generally K-sparse signals, due to the fact that BCH codes are sensitive to noise, we combine a part of syndrome decoding with a compressive sensing-based solver for obtaining $K$ significant frequencies. The computational complexity of our algorithm is \max \left( O(K \log K), O(N) \right), where the Big-O constant of O(N) is very small and only a simple operation involves O(N). Our simulations reveal that O(N) does not dominate the computational cost of sFFT-DT.

cs.DS↗