{"id":21342764,"url":"https://github.com/alex-lechner/carnd-semanticsegmentation-p12","last_synced_at":"2026-04-24T21:33:29.479Z","repository":{"id":159543356,"uuid":"125683435","full_name":"alex-lechner/CarND-SemanticSegmentation-P12","owner":"alex-lechner","description":"An approach for pixel-based classification of drivable parts of the road (known as Semantic Segmentation). Project 12 of Udacity's Self-Driving Car Engineer Nanodegree Program.","archived":false,"fork":false,"pushed_at":"2018-03-19T01:29:21.000Z","size":23267,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"master","last_synced_at":"2025-06-05T07:39:00.354Z","etag":null,"topics":["autonomous-vehicles","deep-learning","deep-neural-networks","deeplearning","fully-convolutional-networks","self-driving-car","semantic-segmentation","tensorflow"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/alex-lechner.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2018-03-18T01:35:54.000Z","updated_at":"2018-09-13T21:35:44.000Z","dependencies_parsed_at":"2023-05-21T13:45:26.310Z","dependency_job_id":null,"html_url":"https://github.com/alex-lechner/CarND-SemanticSegmentation-P12","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/alex-lechner/CarND-SemanticSegmentation-P12","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/alex-lechner%2FCarND-SemanticSegmentation-P12","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/alex-lechner%2FCarND-SemanticSegmentation-P12/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/alex-lechner%2FCarND-SemanticSegmentation-P12/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/alex-lechner%2FCarND-SemanticSegmentation-P12/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/alex-lechner","download_url":"https://codeload.github.com/alex-lechner/CarND-SemanticSegmentation-P12/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/alex-lechner%2FCarND-SemanticSegmentation-P12/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":32241794,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-04-24T13:21:15.438Z","status":"ssl_error","status_checked_at":"2026-04-24T13:21:15.005Z","response_time":64,"last_error":"SSL_read: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["autonomous-vehicles","deep-learning","deep-neural-networks","deeplearning","fully-convolutional-networks","self-driving-car","semantic-segmentation","tensorflow"],"created_at":"2024-11-22T01:09:55.530Z","updated_at":"2026-04-24T21:33:29.465Z","avatar_url":"https://github.com/alex-lechner.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Semantic Segmentation\n\n---\n\n**Semantic Segmentation Project**\n\nThe goals / steps of this project are the following:\n\n* The ``load_vgg``, ``layers``, ``optimize`` \u0026 ``train_nn`` functions are implemented correctly\n* The model decreases loss over time\n* Number of epoch and batch size are set to a reasonable number\n* Most of the images must be labeled correctly (80% of the road and no more than 20% of non-road pixels as the road) \n* Summarize the results with a written report\n\n[//]: # (References)\n[python3]: https://www.python.org/\n[tensorflow]: https://www.tensorflow.org/\n[numpy]: http://www.numpy.org/\n[scipy]: https://www.scipy.org/\n[kitti-download]: http://www.cvlibs.net/download.php?file=data_road.zip\n[starter-code]: https://github.com/udacity/CarND-Semantic-Segmentation\n[img1]: ./imgs/um_000034.png \"Kitti Road Test Image 1\"\n[img2]: ./imgs/umm_000015.png \"Kitti Road Test Image 2\"\n[img3]: ./imgs/uu_000025.png \"Kitti Road Test Image 3\"\n[img4]: ./imgs/umm_000082.png \"Kitti Road Test Image 4\"\n\n---\n\n## Setup\n\n### Frameworks and Packages\nMake sure you have the following installed:\n - [Python 3][python3]\n - [TensorFlow][tensorflow]\n - [NumPy][numpy]\n - [SciPy][scipy]\n\n### Dataset\nIn order to test the Fully Convolutional Network (FCN) download the [Kitti Road dataset][kitti-download]. Extract the dataset in the `data` folder. This will create the folder `data_road` with all the training and test images.\n\n## Files Submitted \u0026 Code Quality\n\n#### 1. The submission includes all required files and every TODO task has been accomplished\nFor this project, I have used the [Semantic Segmentation Starter Code][starter-code] from Udacity and I have modified the following files:\n``main.py``\n\n#### 2. The submission includes functional code\nTo run this project, clone this repository and execute the following command in the repository's root folder:\n```python\npython main.py\n```\n#### 3. Final Output:\nAll labeled test images from the Kitti Road dataset can be found in the ``imgs`` folder of this repository.\nHere are a few examples of the results:\n\n![Kitti Road Test Image 1][img1]\n\n![Kitti Road Test Image 2][img2]\n\n![Kitti Road Test Image 3][img3]\n\n![Kitti Road Test Image 4][img4]\n\n\n## Model Architecture\n\n#### 1. An appropriate model architecture has been employed\nThe model is based on the pre-trained VGG-16 network which was converted to an FCN (Fully Convolutional Network).\n\nFor training the following parameters have been chosen:\n* Epochs = 20\n* Batch size = 18\n* Learning rate = 0.0001\n* Dropout rate = 0.55\n\nThe training loss decreased over time by each epoch. The overall minimum loss was in the last epoch ``Epoch: 20`` and had a ``Training loss: 0.069`` at ``Batch number: 15``.\n\n#### 2. Final Model Architecture\n\nThe final model architecture (``main.py`` lines 63-79) is a Fully Convolutional Network with the following layers and layer sizes:\n\n| Layer                 |     Description                               | \n|:---------------------:|:---------------------------------------------:| \n| Convolution 1x1       | Strides: 1x1                                  |\n| Convolution 1x1       | Strides: 1x1                                  |\n| Convolution 1x1       | Strides: 1x1                                  |\n| Deconvolution 4x4     | Strides: 2x2                                  |\n| Skip Connection layer |                                               |\n| Deconvolution 4x4     | Strides: 2x2                                  |\n| Skip Connection layer |                                               |\n| Deconvolution 16x16   | Strides: 8x8                                  |","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Falex-lechner%2Fcarnd-semanticsegmentation-p12","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Falex-lechner%2Fcarnd-semanticsegmentation-p12","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Falex-lechner%2Fcarnd-semanticsegmentation-p12/lists"}