Media of University of Babylon | Researchers from the College of Information Technology Publish a Scientific Study on Semantic Segmentation Using Liquid Time-Constant Networks with Adaptive Dynamics

Media of University of Babylon - كلية الطب

Researchers from the College of Information Technology Publish a Scientific Study on Semantic Segmentation Using Liquid Time-Constant Networks with Adaptive Dynamics

by : Duhaa Fadill Abbas

Date : 02/6/2026 14:39:05 PM

Veiw : 4

Researchers from the College of Information Technology Publish a Scientific Study on Semantic Segmentation Using Liquid Time-Constant Networks with Adaptive Dynamics

Duhaa Fadill Abbas
As part of the Software Department’s commitment at the College of Information Technology to supporting high-quality scientific research and strengthening the international academic presence of its researchers, PhD candidate Ali Obeid Hassan from the Software Department, under the supervision of Professor Nabil Hashim Al-Araji, has published a specialized scientific study addressing advanced techniques in image semantic segmentation through the utilization of Liquid Time-Constant (LTC) Networks with adaptive dynamics.
The research, entitled “Efficient Semantic Segmentation via Liquid Time-Constant Networks with Adaptive Dynamics,” was published in the International Journal of Intelligent Networks and Systems Society (INASS), a Scopus-indexed Q2 journal, reflecting the scientific significance and academic value of the study.
The research presents an advanced approach to semantic image segmentation by integrating Liquid Time-Constant (LTC) Networks with conventional Convolutional Neural Networks (CNNs), aiming to achieve competitive performance while reducing model size and computational complexity.
These networks are inspired by biological neural circuits, enabling adaptive processing dynamics that respond to the characteristics of input data. The proposed framework employs an encoder–decoder architecture enhanced with a Closed-form Continuous-time (CfC) feature adaptation head, allowing spatial features to be processed through temporal dynamics inspired by biological systems and providing a more efficient alternative to traditional approaches.
The study aims to support the development of intelligent computer vision systems that combine efficiency, adaptability, and scalability, making them suitable for deployment in resource-constrained environments such as unmanned aerial vehicles (UAVs), mobile devices, and embedded systems. The proposed approach contributes to advancing research and applications in the fields of computer vision, scene understanding, and intelligent image analysis..

Research Link :
https://inass.org/wp-content/uploads/2026/01/2026063009-2.pdf