Update README.md

README.md

@@ -1,14 +1,11 @@
 ---
-language:
-- en
 pipeline_tag: object-detection
-tags:
-- code
 ---
 # MSRNet: A Multi-Scale Recursive Network for Camouflaged Object Detection.
 This repository is for MSRNet: A Multi-Scale Recursive Network for Camouflaged Object Detection, introduced in this [paper](https://huggingface.co/papers/2511.12810).
 
-The model is built using PyTorch on an NVIDIA RTX A6000 GPU with a total memory size of 48 GB. The code is also available in our [GitHub](https://github.com/linaagh98/MSRNet) repository.
+The model is built using PyTorch on an NVIDIA RTX A6000 GPU with a total memory size of 48 GB.
+The code is also available in our [GitHub](https://github.com/linaagh98/MSRNet) repository.
 
 ## Contents
 1. [Introduction](#introduction)
@@ -21,11 +18,12 @@ The model is built using PyTorch on an NVIDIA RTX A6000 GPU with a total memory
 9. [Acknowledgement](#acknowledgment)
 
 ## Introduction
-We introduce a Multi-Scale Recursive Network that utilizes a Pyramid Vision Transformer backbone to extract multi-scale features. This network employs Attention-Based Scale Integration Units for selective feature merging, and a recursive decoding strategy incorporating Multi-Granularity Fusion Units to refine features and enhance global context. Our approach leverages multi-scale learning and recursive optimization, achieving state-of-the-art performance on benchmark datasets for detecting small and multiple camouflaged objects.
+Camouflaged object detection (COD) is a challenging computer vision task that aims to identify and segment objects blending into their environment. MSRNet proposes a Multi-Scale Recursive Network to extract and combine multi-scale features, using a Pyramid Vision Transformer backbone and Attention-Based Scale Integration Units. Its decoder refines features recursively with Multi-Granularity Fusion Units and a novel recursive-feedback decoding strategy to enhance global context. This approach improves detection of small and multiple camouflaged objects, achieving state-of-the-art results on several benchmark datasets.
+
 ## Network
 This diagram illustrates the overall architecture of MSRNet. 
 ![Methodology](Images/MethodologyDiagram.png)
-
+![Methodology](https://raw.githubusercontent.com/linaagh98/MSRNet/main/Images/MethodologyDiagram.png)
 
 ## Data Preparation
 In this research, we utilized four benchmark datasets for camouflaged object detection (CAMO, CHAMELEON, COD10K, and NC4K).