Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Mahmoud Tahmasebi has a total of 11 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    ESMStereo: Enhanced ShuffleMixer Disparity Upsampling for Real-Time and Accurate Stereo Matching

    by Mahmoud Tahmasebi, Saif Huq, Kevin Meehan, Marion McAfee

    Stereo matching has become an increasingly important component of modern autonomous systems. Developing deep learning-based stereo-matching models that deliver high accuracy while operating in real time continues to be a major challenge in computer vision. In the domain of cost volume-based stereo matching, accurate disparity estimation depends heavily on large-scale cost volumes. However, such large volumes store substantial redundant information and also require computationally intensive aggregation units for processing and regression, making real-time performance unattainable. Conversely, small-scale cost volumes followed by lightweight aggregation units provide a promising route for real-time performance, but lack sufficient information to ensure highly accurate disparity estimation. To address this challenge, we propose the Enhanced Shuffle Mixer (ESM) to mitigate information loss associated with small-scale cost volumes. ESM restores critical details by integrating primary features into the disparity upsampling unit. It quickly extracts features from the initial disparity estimation and fuses them with image features. These features are mixed by shuffling and layer splitting, then refined through a compact feature-guided hourglass network to recover more detailed scene geometry. The ESM focuses on local contextual connectivity with a large receptive field and low computational cost, leading to improved disparity estimation accuracy while maintaining real-time performance under the evaluated settings. The compact version of ESMStereo achieves an inference speed of 116 FPS on RTX 4070S and 91 FPS on the AGX Orin.

  • Enhancing Stereo Matching with Multi-Scale Selective Attention Aggregation

    by Mahmoud Tahmasebi, Jack Quinn, Jamie Roche, Saif Huq, Kevin Meehan, Marion McAfee

    Stereo matching is a fundamental task in computer vision, enabling applications such as robotics, autonomous systems, and augmented reality to gain depth understanding. Currently, the design and development of an accurate and efficient stereo matching model have remained an active area of research with significant potential for advancement. In this work, we introduce a high-performance stereo matching model that improves the conventional cost-volume-based pipeline by integrating a Selective Attention (SA) module into the 3D aggregation block. SA leverages a dual-branch structure with two distinct kernel sizes to capture multi-scale contextual information. It effectively fuses fine-grained spatial details from initial features with broader contextual cues in the 3D aggregation block. The module employs a dynamic selection mechanism to adjust kernel weights adaptively, enhancing the model’s capability to handle occlusions, textureless regions, and varying object scales. The source code will be publicly available at https://github.com/M2219/StereoSA.

  • Open Access

    ESMStereo: Enhanced ShuffleMixer Disparity Upsampling for Real-Time and Accurate Stereo Matching

    by Mahmoud Tahmasebi, Saif Huq, Kevin Meehan, Marion McAfee

    Stereo matching has become an increasingly important component of modern autonomous systems. Developing deep learning-based stereo matching models that deliver high accuracy while operating in real-time continues to be a major challenge in computer vision. In the domain of cost-volume-based stereo matching, accurate disparity estimation depends heavily on large-scale cost volumes. However, such large volumes store substantial redundant information and also require computationally intensive aggregation units for processing and regression, making real-time performance unattainable. Conversely, small-scale cost volumes followed by lightweight aggregation units provide a promising route for real-time performance, but lack sufficient information to ensure highly accurate disparity estimation. To address this challenge, we propose the Enhanced Shuffle Mixer (ESM) to mitigate information loss associated with small-scale cost volumes. ESM restores critical details by integrating primary features into the disparity upsampling unit. It quickly extracts features from the initial disparity estimation and fuses them with image features. These features are mixed by shuffling and layer splitting then refined through a compact feature-guided hourglass network to recover more detailed scene geometry. The ESM focuses on local contextual connectivity with a large receptive field and low computational cost, leading to the reconstruction of a highly accurate disparity map at real-time. The compact version of ESMStereo achieves an inference speed of 116 FPS on high-end GPUs and 91 FPS on the AGX Orin.

  • Open Access

    DCVSMNet: Double Cost Volume Stereo Matching Network

    by Mahmoud Tahmasebi, Saif Huq, Kevin Meehan, Marion McAfee


  • Dynamic Mode Decomposition and Koopman Operator for Iterated Function Systems

    by Ramen Ghosh, Mahmoud Tahmasebi, Marion McAfee

    The article aims to demonstrate the effectiveness of using the Koopman operator and dynamic mode decomposition (DMD) for iterated function systems (IFS). Specifically, we show how these tools can be used to analyze and predict the behaviour of stochastic nonlinear dynamical systems represented by discrete-time Markov chains on a compact state space. In particular, we focus on an ergodic nonlinear IFS, which has not been studied with these approaches before. Our paper presents the first application of dynamic mode decomposition to this type of system.