The paper introduces the Multi-level Multi-scale Gated Attentional Squeezed Network (2MGAS-Net), a deep learning model designed for accurate polyp segmentation in endoscopic images. This model addresses the challenges of polyp characteristics such as size, color, shape, position, and complex structures. The key contributions include: 1. **2MGAS-Net Framework**: A novel deep learning architecture that integrates Multi-scale Gated Attentional Squeezed Feature Fusion (MGAS2F) units to enhance polyp segmentation, particularly for small and crucial lesion boundaries. 2. **MGAS2F Strategy**: Combines Multi-scale Squeezed Feature Fusion (MS2F) and Cascaded Gated Attentional Transformer (CGA-T) modules to capture contextual information at multiple scales. 3. **CGA-T Module**: Enhances the model's ability to estimate polyp boundaries accurately. 4. **Performance Evaluation**: Demonstrates superior performance compared to existing state-of-the-art methods on publicly available datasets, indicating its potential to improve clinical decision-making and diagnostic approaches for colorectal cancer. The paper also reviews related works, highlighting advancements in automatic segmentation algorithms, attention mechanisms, and multi-level multi-scale feature extraction approaches, and discusses the limitations of current methods.The paper introduces the Multi-level Multi-scale Gated Attentional Squeezed Network (2MGAS-Net), a deep learning model designed for accurate polyp segmentation in endoscopic images. This model addresses the challenges of polyp characteristics such as size, color, shape, position, and complex structures. The key contributions include: 1. **2MGAS-Net Framework**: A novel deep learning architecture that integrates Multi-scale Gated Attentional Squeezed Feature Fusion (MGAS2F) units to enhance polyp segmentation, particularly for small and crucial lesion boundaries. 2. **MGAS2F Strategy**: Combines Multi-scale Squeezed Feature Fusion (MS2F) and Cascaded Gated Attentional Transformer (CGA-T) modules to capture contextual information at multiple scales. 3. **CGA-T Module**: Enhances the model's ability to estimate polyp boundaries accurately. 4. **Performance Evaluation**: Demonstrates superior performance compared to existing state-of-the-art methods on publicly available datasets, indicating its potential to improve clinical decision-making and diagnostic approaches for colorectal cancer. The paper also reviews related works, highlighting advancements in automatic segmentation algorithms, attention mechanisms, and multi-level multi-scale feature extraction approaches, and discusses the limitations of current methods.