Sundaram Muthu

Dr. Sundaram Muthu


Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
7 April 2020 in IEEE Transactions on Image Processing
Verified
Journal Article
Closed access
SMSundaram Muthu
Sundaram Muthu
RTRuwan Tennakoon
Ruwan Tennakoon
TRTharindu Rathnayake
Tharindu Rathnayake
Reza HosseinnezhadReza Hosseinnezhad
Reza Hosseinnezhad
DSDavid Suter
David Suter
ABAlireza Bab-Hadiashar
Alireza Bab-Hadiashar
This paper presents an innovative method for motion segmentation in RGB-D dynamic videos with multiple moving objects. The focus is on finding static, small or slow moving objects (often overlooked by other methods) that their inclusion can improve the motion segmentation results. In our approach, semantic object based segmentation and motion cues are combined to estimate the number of moving objects, their motion parameters and perform segmentation. Selective object-based sampling and correspondence matching are used to estimate object specific motion parameters. The main issue with such an approach is the over segmentation of moving parts due to the fact that different objects can have the same motion (e.g. background objects). To resolve this issue, we propose to identify objects with similar motions by characterizing each motion by a distribution of a simple metric and using a statistical inference theory to assess their similarities. To demonstrate the significance of the proposed statistical inference, we present an ablation study, with and without static objects inclusion, on SLAM accuracy using the TUM-RGBD dataset. To test the effectiveness of the proposed method for finding small or slow moving objects, we applied the method to RGB-D MultiBody and SBM-RGBD motion segmentation datasets. The results showed that we can improve the accuracy of motion segmentation for small objects while remaining competitive on overall measures.

5 February 2025 in Sensors
Verified
Journal Article
Open access
PDF file
FMFahira Afzal Maken
Fahira Afzal Maken
SMSundaram Muthu
Sundaram Muthu
CNChuong Nguyen
Chuong Nguyen
CSChangming Sun
Changming Sun
JTJinguang Tong
Jinguang Tong
LPLars Petersson
Lars Petersson
High-resolution RGB-D sensors are widely used in computer vision, manufacturing, and robotics. The depth maps from these sensors have inherently high measurement uncertainty that includes both systematic and non-systematic noise. These noisy depth estimates degrade the quality of scans, resulting in less accurate 3D reconstruction, making them unsuitable for some high-precision applications. In this paper, we focus on quantifying the uncertainty in the depth maps of high-resolution RGB-D sensors for the purpose of improving 3D reconstruction accuracy. To this end, we estimate the noise model for a recent high-precision RGB-D structured light sensor called Zivid when mounted on a robot arm. Our proposed noise model takes into account the measurement distance and angle between the sensor and the measured surface. We additionally analyze the effect of background light, exposure time, and the number of captures on the quality of the depth maps obtained. Our noise model seamlessly integrates with well-known classical and modern neural rendering-based algorithms, from KinectFusion to Point-SLAM methods using bilinear interpolation as well as 3D analytical functions. We collect a high-resolution RGB-D dataset and apply our noise model to improve tracking and produce higher-resolution 3D models.

10 August 2022 in International Journal of Machine Learning and Cybernetics
Verified
Journal Article
Open access
PDF file
SMSundaram Muthu
Sundaram Muthu
RTRuwan Tennakoon
Ruwan Tennakoon
RHReza Hoseinnezhad
Reza Hoseinnezhad
ABAlireza Bab-Hadiashar
Alireza Bab-Hadiashar
This paper presents a new approach, called TMNet, to solve unsupervised video object segmentation (UVOS) problem. The UVOS is still a challenging problem as prior methods suffer from issues like generalization errors in unseen test videos, over reliance on optic flow, and capturing fine details at object boundaries. These issues make the UVOS an ill-defined problem, particularly in presence of multiple objects. Our focus is to constrain the problem and improve the segmentation results by fusion of multiple available cues such as appearance and motion, as well as image and flow edges. To constrain the problem, instead of predicting segmentation directly, we predict affinities between neighbouring pixels for being part of the same object and cluster those to obtain category agnostic segmentation. To further improve the segmentation, we fuse multiple-sources of information through a novel Temporal Motion Attention (TMA) module that uses neural attention to learn powerful spatio-temporal features. In addition, we also design an edge refinement module (using image and optic flow edges) to refine and improve the accuracy of object segmentation boundaries. The overall framework is capable of segmenting and finding accurate objects’ boundaries without any heuristic post processing. This enables the method to be used for unseen videos. Experimental results on challenging DAVIS16 and multi object DAVIS17 datasets show that our proposed TMNet performs favorably compared to the state-of-the-art methods without post processing.

11 September 2023 in IEEE Access
Verified
Journal Article
Open access
SMSundaram Muthu
Sundaram Muthu
RTRuwan Tennakoon
Ruwan Tennakoon
Reza HosseinnezhadReza Hosseinnezhad
Reza Hosseinnezhad
ABAlireza Bab-Hadiashar
Alireza Bab-Hadiashar
The analysis of 3D motion information is the key to solve various computer vision tasks. Scene flow estimation tackles the problem of obtaining the 3D motion field. In this paper, we review the recent scene flow estimation papers with a focus on learning-based methods. The problem formulation, challenges and applications are introduced. The existing datasets and performance metrics are presented. The reason behind learning-based methods replacing the traditional variational methods are discussed. CNN-based scene flow estimation methods are then categorized with respect to the level of supervision, data-availability and the number of steps involved in obtaining the results. The performance of different methods on the well known KITTI and FlyingThings3D datasets are tabulated. Their relative advantages and limitations are then analysed. Future trends and some open problems in the estimation of scene flow are discussed with special focus on the self-supervised methods that does not require labelled training data.

19 August 2023 in Computer Vision and Image Understanding
Verified
Journal Article
Open access
SMSundaram Muthu
Sundaram Muthu
RTRuwan Tennakoon
Ruwan Tennakoon
TRTharindu Rathnayake
Tharindu Rathnayake
RHReza Hoseinnezhad
Reza Hoseinnezhad
DSDavid Suter
David Suter
ABAlireza Bab-Hadiashar
Alireza Bab-Hadiashar

24 March 2023 in arXiv
Verified
Preprint
Open access
PDF file
JTJinguang Tong
Jinguang Tong
SMSundaram Muthu
Sundaram Muthu
FMFahira Afzal Maken
Fahira Afzal Maken
CNChuong Nguyen
Chuong Nguyen
HLHongdong Li
Hongdong Li
In this paper, we define a new problem of recovering the 3D geometry of an object confined in a transparent enclosure. We also propose a novel method for solving this challenging problem. Transparent enclosures pose challenges of multiple light reflections and refractions at the interface between different propagation media e.g. air or glass. These multiple reflections and refractions cause serious image distortions which invalidate the single viewpoint assumption. Hence the 3D geometry of such objects cannot be reliably reconstructed using existing methods, such as traditional structure from motion or modern neural reconstruction methods. We solve this problem by explicitly modeling the scene as two distinct sub-spaces, inside and outside the transparent enclosure. We use an existing neural reconstruction method (NeuS) that implicitly represents the geometry and appearance of the inner subspace. In order to account for complex light interactions, we develop a hybrid rendering strategy that combines volume rendering with ray tracing. We then recover the underlying geometry and appearance of the model by minimizing the difference between the real and hybrid rendered images. We evaluate our method on both synthetic and real data. Experiment results show that our method outperforms the state-of-the-art (SOTA) methods. Codes and data will be available at https://github.com/hirotong/ReNeuS