{"id":19121826,"url":"https://github.com/bshao001/dmsmsgrcg","last_synced_at":"2025-05-05T16:28:41.491Z","repository":{"id":201721666,"uuid":"92323906","full_name":"bshao001/DmsMsgRcg","owner":"bshao001","description":"A photo OCR project aims to output DMS messages contained in sign structure images.","archived":false,"fork":false,"pushed_at":"2017-12-14T00:28:58.000Z","size":111,"stargazers_count":18,"open_issues_count":0,"forks_count":10,"subscribers_count":2,"default_branch":"master","last_synced_at":"2025-04-19T10:28:28.176Z","etag":null,"topics":["convolutional-neural-networks","image-classification","image-processing","image-recognition","keras","object-detection","ocr","photo-ocr","tensorflow","yolo"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/bshao001.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE.txt","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null}},"created_at":"2017-05-24T18:28:11.000Z","updated_at":"2019-09-25T08:15:47.000Z","dependencies_parsed_at":null,"dependency_job_id":"f86fc136-ded2-45da-8a03-6c9646c59b68","html_url":"https://github.com/bshao001/DmsMsgRcg","commit_stats":null,"previous_names":["bshao001/dmsmsgrcg"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bshao001%2FDmsMsgRcg","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bshao001%2FDmsMsgRcg/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bshao001%2FDmsMsgRcg/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bshao001%2FDmsMsgRcg/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/bshao001","download_url":"https://codeload.github.com/bshao001/DmsMsgRcg/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":252533384,"owners_count":21763588,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","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":["convolutional-neural-networks","image-classification","image-processing","image-recognition","keras","object-detection","ocr","photo-ocr","tensorflow","yolo"],"created_at":"2024-11-09T05:18:38.244Z","updated_at":"2025-05-05T16:28:41.460Z","avatar_url":"https://github.com/bshao001.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# DmsMsgRcg\n\n![](https://img.shields.io/badge/python-3.6.2-brightgreen.svg)  ![](https://img.shields.io/badge/tensorflow-1.4.0-yellowgreen.svg?sanitize=true)\n\nA photo OCR project aims to recognize and output DMS messages contained in sign structure images.\n\n## Project Details\nThis project will provide an aided function to a well-established highway management software - Operations Task Manager \n(OTM) used in FDOT. The code is and will be implemented in Python with TensorFlow (and tf.keras in TF 1.4). The images\nin this project are all of the same size (height * width = 480 * 640) and same format (in JPG). By design, the pipeline \nwill include the following 4 phases or steps:\n\n#### 1. Text area detection \nIt is like object detection. This step try to locate where the text areas are. YOLO algorithm is slightly modified to \nadapt the need of this step. It is possible to know how many message areas are when feeding a sign structure image. \nThere are also hint information available about whether the message areas contain Toll information (which are mainly \ndigits) or status information (which mostly be 26 upper-case English letters). However, the image may contain other \ntexts that can are same or similar words that should be discarded, for example, words - \u003cem\u003eEXPRESS LANES\u003c/em\u003e or \ndigits might appear in a text area we care or appear in the text area where we should ignore, with very similar fonts. \n\nYou can find a sliding window version implementation for this step in the other branch in this repository. The training \nexamples are generated both manually and by using the old text detector. A python script to create labels is included. \nIt requires bounding boxes (with pure red, i.e. RGB = 255, 0, 0 and thickness = 1) be drawn on the images, which are then\nsave in PNG format.\n\nThe code is completed for this step.\n\n#### 2. Text type classification \nThis extra step tries to classify the cropped images from the first step into their corresponding types. This will rule \nout the need to recognize complicated Lane Status Messages.\n\nDue to the new introduction of TensorFlow Keras, this step and the following two steps are under development.\n\n#### 3. Character segmentation \nThis will split the input image, which contains all the message characters, into individual character images.\n\n#### 4. Character recognition\nIn this stage, we will output a single character for each input image.\n\nFinally, all the information gets re-organized.\n\nThere are training and prediction associated with each stage. Theoretically, step 1 and 2 can be combined, however, \nthat would require much more training examples, and the labelling process would be much more complicated.\n\nDue to copyright issue, the training images cannot be exposed here.\n\n## References:\n1. https://github.com/experiencor/basic-yolo-keras\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbshao001%2Fdmsmsgrcg","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fbshao001%2Fdmsmsgrcg","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbshao001%2Fdmsmsgrcg/lists"}