Update 2024-03-22-Week-3-4.md

docs/_posts/2024-03-22-Week-3-4.md

@@ -1,53 +1,52 @@
 ---
-title: "Week 2 - Semantic segmentation and Github Pages"
+title: "Week 3-4 - Semantic segmentation and Github Pages II"
 image: ../assets/images/logo.png
 categories:
   - Weekly Log
 tags:
   - Semantic segmentation
   - Github pages
-  - Cityscapes metrics
 
 ---
 
 # Objectives
 
-1. Continue adapting DeepLabV3+ with the RELLIS-3D and RUGD datasets.
-2. GitHub Pages.
-3. Perform Semantic Segmentation on Cityscapes dataset.
-4. Study the metrics and models proposed on the Cityscapes dataset website: https://www.cityscapes-dataset.com/benchmarks/
+1. Continue adapting Ganav and DeepLabV3+ with the RELLIS-3D and RUGD datasets.
+2. GitHub Pages cloning repository.
+3. Create post for the current and the last 3 weeks work progress.
 
 ---
 
 # Progress This Week
 
-## 1. DeepLabV3+ on Rellis-3D 
+## 1. GANav and DeepLabV3+ on Rellis-3D 
 
-This week I have had different problems. I have not been able to adapt the DeepLabV3 model with the RELLIS-3D dataset, it requires image preprocessing that has not been applied. On the other hand, I have not been able to successfully create the Github pages sections.
+Reproduction the GANav-offroad Github repository to perform semantic segmentation with the RELLIS-3D and RUGD datasets. RUGD have 24 semantic categories and RELLIS-3D have 20 semantic categories. The GANav-offroad repository reprocesses the two datasets to group the categories into 6 groups. To do this, new mask images are created with the new category ids and these are used for semantic segmentation. The repository was reproduced on a local computer and the code was executed without any modification.
 
-## 2. Github Pages
+The adaptation of Rellis-3D to DeepLabV3 could not be achieved, I still have problems adapting the data set to the network. 
+The reasons are still being investigated.
 
-During today's meeting I met Sergio who will help me in the development of the TFM. He was telling me how to manage Github pages to clone the "docs" folder and be able to make weekly posts.
+## 2. Github Pages
 
-## 3. Cityscapes Semantic segmentation 
+The Github Pages template could be cloned.
 
-Performed semantic segmentation with DeepLabV3 on Cityscapes dataset.
+## 3. Create post for the current and the last 3 weeks work progress.
 
-## 4. Cityscapes metrics
+The content from previous weeks could be updated in new posts.
 
-In different studies that compare semantic segmentation models that I have been able to read (among them cityscapes), in addition to the conventional metrics (F1-score, TP, TN, FP, FN), the IoU (Interception over Union) metrics are usually used, both for classes as categories.
 
 ---
 
 # Next Week Work Planning
 
-  1. Continue adapting DeepLabV3+ and GanAV with the RELLIS-3D and RUGD datasets.
-  2. GitHub Pages cloning repository.
-  3. Create post for the current and the last 3 weeks work progress. 
+  1. Continue adapting DeepLabV3+ with the RELLIS-3D dataset.
+  2. Update Github Pages.
 
 
 ---
 
 # References
 
-* [Cityscapes Metrics] https://www.cityscapes-dataset.com/benchmarks/#instance-level-scene-labeling-task
+* [Rellis-3D] https://www.unmannedlab.org/research/RELLIS-3D
+* [RUGD Dataset] http://rugd.vision/
+* [GANav-offroad] https://github.com/rayguan97/GANav-offroad