{"id":2,"date":"2022-01-07T14:12:52","date_gmt":"2022-01-07T14:12:52","guid":{"rendered":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/?page_id=2"},"modified":"2022-12-20T16:10:53","modified_gmt":"2022-12-20T16:10:53","slug":"experimental-setup","status":"publish","type":"page","link":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/experimental-setup\/","title":{"rendered":"Experimental Setup"},"content":{"rendered":"\n<div class=\"wp-block-group\"><div class=\"wp-block-group__inner-container is-layout-flow wp-block-group-is-layout-flow\">\n<h2 class=\"wp-block-heading\"><strong>Datasets<\/strong><\/h2>\n\n\n\n<p>As mentioned earlier, there is no well-defined work zone targeted data available and gaining access to work-zone relevant data to train the models has been the main focus of our efforts. This can be achieved in two ways:<\/p>\n\n\n\n<ul class=\"wp-block-list\"><li><a href=\"#app1\"><mark class=\"has-inline-color has-vivid-purple-color\">Approach 1<\/mark><\/a>: Exploring existing datasets and models trained on them with at least a sub set of the classes we are interested in.<\/li><li><a href=\"#app2\"><mark class=\"has-inline-color has-vivid-purple-color\">Approach 2<\/mark><\/a>: Collecting our own new data and annotating it.<\/li><li><a href=\"#app3\"><mark class=\"has-inline-color has-vivid-purple-color\">Approach 3<\/mark><\/a>: Augment the data to generate more data.<\/li><\/ul>\n\n\n\n<div id=\"app1\"><h2>Approach 1: <strong>Explore Existing Datasets<\/strong><\/h2><\/div>\n\n\n\n<p>As part of the early efforts in this project, we tried exploring existing datasets which had some representation for what we considered as &#8220;relevant&#8221; in the context of work-zones.<\/p>\n\n\n\n<p>They include the following datasets:<\/p>\n\n\n\n<ul class=\"wp-block-list\"><li>ACID<\/li><li>Roadbotics<\/li><li>LVIS<\/li><li>NuScenes<\/li><li>Argoverse 2.0<\/li><\/ul>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>LVIS Dataset<\/strong><\/h3>\n\n\n\n<p>We explored the Large Vocabulary Instance Segmentation (LVIS) Dataset which is a long tailed dataset with a couple relevant classes to construction zones:<\/p>\n\n\n\n<div class=\"wp-block-image\"><figure class=\"aligncenter size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"485\" height=\"556\" src=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.22.16-PM.png\" alt=\"\" class=\"wp-image-70\" srcset=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.22.16-PM.png 485w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.22.16-PM-262x300.png 262w\" sizes=\"auto, (max-width: 485px) 100vw, 485px\" \/><\/figure><\/div>\n\n\n\n<p>We found 25 relevant classes as mentioned below. <\/p>\n\n\n\n<div class=\"wp-block-image\"><figure class=\"aligncenter size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"475\" src=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.24.02-PM-1024x475.png\" alt=\"\" class=\"wp-image-73\" srcset=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.24.02-PM-1024x475.png 1024w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.24.02-PM-300x139.png 300w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.24.02-PM-768x356.png 768w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.24.02-PM.png 1199w\" sizes=\"auto, (max-width: 767px) 89vw, (max-width: 1000px) 54vw, (max-width: 1071px) 543px, 580px\" \/><\/figure><\/div>\n\n\n\n<p>We incorporated LVIS Detection along with COCO Detections already running on static cameras installed at Pittsburgh Intersections and are currently investigating the data. Any time a class from these 25 seemingly relevant classes are detected, we go ahead and sample those video frames to check if there are construction zones. <\/p>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>NuScenes Dataset<\/strong><\/h3>\n\n\n\n<p>LVIS has some relevant classes but misses out on two of the most important classes: Construction Workers and Heavy Machines. NuScenes has these two classes and some other construction zone relevant classes. We found 5 classes of relevance and are exploring those.<\/p>\n\n\n\n<div class=\"wp-block-image\"><figure class=\"aligncenter size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"568\" height=\"416\" src=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.29.23-PM.png\" alt=\"\" class=\"wp-image-79\" srcset=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.29.23-PM.png 568w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.29.23-PM-300x220.png 300w\" sizes=\"auto, (max-width: 568px) 100vw, 568px\" \/><\/figure><\/div>\n\n\n\n<p>We the trained a NuScenes-Construction Object Detector which is trained on only these 5 classes and can help us detect construction zones with better confidence.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>Argoverse 2.0<\/strong><\/h3>\n\n\n\n<p>We also explored the recently released Argoverse 2.0 &#8211; Sensor Dataset for any construction zone relevant classes. We found a couple relevant classes.<\/p>\n\n\n\n<div class=\"wp-block-image\"><figure class=\"aligncenter size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"397\" src=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.30.30-PM-1024x397.png\" alt=\"\" class=\"wp-image-83\" srcset=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.30.30-PM-1024x397.png 1024w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.30.30-PM-300x116.png 300w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.30.30-PM-768x298.png 768w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.30.30-PM.png 1116w\" sizes=\"auto, (max-width: 767px) 89vw, (max-width: 1000px) 54vw, (max-width: 1071px) 543px, 580px\" \/><\/figure><\/div>\n<\/div><\/div>\n\n\n\n<p>This way, we tried training our Faster R-CNN detection network using these datasets to be able to at least identify those image frames with some work-zone relevant classes.<\/p>\n\n\n\n<p>The training and testing pipeline for this approach, looked as follows:<\/p>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>Training Pipeline<\/strong><\/h2>\n\n\n\n<div class=\"wp-block-image\"><figure class=\"aligncenter size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"392\" src=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.44.53-PM-1024x392.png\" alt=\"\" class=\"wp-image-93\" srcset=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.44.53-PM-1024x392.png 1024w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.44.53-PM-300x115.png 300w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.44.53-PM-768x294.png 768w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.44.53-PM.png 1320w\" sizes=\"auto, (max-width: 767px) 89vw, (max-width: 1000px) 54vw, (max-width: 1071px) 543px, 580px\" \/><\/figure><\/div>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>Testing Pipeline<\/strong><\/h2>\n\n\n\n<h3 class=\"has-white-color has-text-color wp-block-heading\"><\/h3>\n\n\n\n<div class=\"wp-block-image\"><figure class=\"aligncenter size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"445\" src=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/old_test-1024x445.jpeg\" alt=\"\" class=\"wp-image-235\" title=\"Training Pipeline\" srcset=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/old_test-1024x445.jpeg 1024w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/old_test-300x130.jpeg 300w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/old_test-768x334.jpeg 768w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/old_test-1536x667.jpeg 1536w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/old_test.jpeg 1600w\" sizes=\"auto, (max-width: 767px) 89vw, (max-width: 1000px) 54vw, (max-width: 1071px) 543px, 580px\" \/><figcaption><em>Test Pipeline <sup>[1]<\/sup><\/em><br><\/figcaption><\/figure><\/div>\n\n\n\n<p>We feed in surveillance videos captured by the cameras at intersections. Sample 120 frames every 10 minutes, perform median subtraction and if we do find difference, we perform detection for COCO and the selected &#8216;work-zone&#8217; relevant LVIS, NuScenes and ArgoVerse 2.0 classes and then perform further analysis on top of the detection results to finally pick frames that contain work-zone scenes in them.<\/p>\n\n\n\n<div id=\"app2\">\n<h2>Approach 2: <strong>Collect and Annotate to Generate a new Dataset<\/strong><\/h2><\/div>\n\n\n\n<p>Post incorporating existing datasets and analyzing performance for construction zone relevant categories, it was realized that more construction zone relevant categories with more data per category were required for successfully calling out construction zones. The main reason being, the existing datasets are long-tailed and not targeted towards the task of construction zone detection and hence lack variety and enough work-zone relevant class representation in them. Therefore ILIM Lab at CMU undertook a data collection and annotation effort to get object bounding boxes and segmentation masks for 33 construction zone relevant object categories. The effort resulted in total of <strong>2885 images<\/strong>, <strong>30202 annotations<\/strong> and <strong>33 categories<\/strong>. The statistics are depicted below.<\/p>\n\n\n\n<p>PS: The annotation effort is ongoing and we expect to get access to more data in due course.<\/p>\n\n\n\n<p><img decoding=\"async\" width=\"289px;\" height=\"431px;\" src=\"https:\/\/lh5.googleusercontent.com\/6uPPl1SQ5WaLSwmXlX76vxEFl_5stl2tZ6napy0nNVgXUveuxEffTXAkyXdKKo72JduCEoC9-QLxsFf0gSePYyQcYIVL9atQGuTY5uzgKFNclYeELYs2vJF86TvyALUJL1A0Ad-QdgEAdkdp4wZnhmsY8g=s2048\"><\/p>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>Revised Training Pipeline<\/strong><\/h2>\n\n\n\n<p>Image segmentation annotations were acquired from the data annotation effort. The training pipeline was augmented to predict segmentation masks using fine-tuning of MaskRCNN. The detection-cum-segmentation results were utilized to perform probabilistic analysis for presence of construction zones. <\/p>\n\n\n\n<figure class=\"wp-block-image size-large is-style-rounded\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"413\" src=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/Screen-Shot-2022-12-17-at-1.07.05-PM-1024x413.png\" alt=\"\" class=\"wp-image-246\" srcset=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/Screen-Shot-2022-12-17-at-1.07.05-PM-1024x413.png 1024w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/Screen-Shot-2022-12-17-at-1.07.05-PM-300x121.png 300w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/Screen-Shot-2022-12-17-at-1.07.05-PM-768x310.png 768w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/Screen-Shot-2022-12-17-at-1.07.05-PM.png 1304w\" sizes=\"auto, (max-width: 767px) 89vw, (max-width: 1000px) 54vw, (max-width: 1071px) 543px, 580px\" \/><\/figure>\n\n\n\n<div id=\"app3\"><h2>Approach 3:<strong> Augment Data<\/strong><\/h2><\/div>\n\n\n\n<p>In an effort to add more data to the pipeline, 2D and 3D data annotation methods are explored.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>2D Clip-And-Paste Data Augmentation<\/strong><\/h2>\n\n\n\n<p>For 2D augmentations, cutting and pasting technique was used on 4 inter-changeable categories &#8211; Cones, drums, vertical panels and TTC message boards. They are similar in respect to the their aspect ratios and hence were randomly swapped with each other in the training dataset to augment and boost data in these 4 categories. <\/p>\n\n\n\n<figure class=\"wp-block-image size-large is-resized is-style-default\"><img loading=\"lazy\" decoding=\"async\" src=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/2d_clip-n-paste-1024x486.jpeg\" alt=\"\" class=\"wp-image-201\" width=\"674\" height=\"319\" srcset=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/2d_clip-n-paste-1024x486.jpeg 1024w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/2d_clip-n-paste-300x142.jpeg 300w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/2d_clip-n-paste-768x364.jpeg 768w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/2d_clip-n-paste-1536x729.jpeg 1536w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/2d_clip-n-paste.jpeg 1600w\" sizes=\"auto, (max-width: 674px) 100vw, 674px\" \/><\/figure>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>Mixed-Reality based 3D Synthetic Data Augmentation<\/strong><\/h2>\n\n\n\n<p>We collected work-zone relevant 3D assets like JCB, cones etc. and performed a mixed reality based 3D-Synthetic data augmentation using SfM. We use a base image we intend to augment and query Google Street View API to get access to Street View images in its vicinity and perform sparse view reconstruction of the scene. Using the relative pose information between the cameras, we then compute the respective transforms and render the 3D asset from realistic view points and there by generate geometrically consistent multi-view synthetic data with precise 2D masks and annotations. The main limitation with this approach was to get access to enough work-zone relevant 3D assets.<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"420\" src=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/WhatsApp-Image-2022-12-17-at-12.56.53-1024x420.jpeg\" alt=\"\" class=\"wp-image-238\" srcset=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/WhatsApp-Image-2022-12-17-at-12.56.53-1024x420.jpeg 1024w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/WhatsApp-Image-2022-12-17-at-12.56.53-300x123.jpeg 300w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/WhatsApp-Image-2022-12-17-at-12.56.53-768x315.jpeg 768w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/WhatsApp-Image-2022-12-17-at-12.56.53-1536x630.jpeg 1536w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/WhatsApp-Image-2022-12-17-at-12.56.53.jpeg 1600w\" sizes=\"auto, (max-width: 767px) 89vw, (max-width: 1000px) 54vw, (max-width: 1071px) 543px, 580px\" \/><figcaption>Mixed-Reality based 3D Synthetic Data Augmentation Pipeline<\/figcaption><\/figure>\n\n\n\n<p>The testing pipeline used was similar to the previous setup.<\/p>\n\n\n\n<figure class=\"wp-block-image size-large is-style-rounded\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"454\" src=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/new_test-1024x454.jpeg\" alt=\"\" class=\"wp-image-250\" srcset=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/new_test-1024x454.jpeg 1024w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/new_test-300x133.jpeg 300w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/new_test-768x341.jpeg 768w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/new_test-1536x681.jpeg 1536w, https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/new_test.jpeg 1585w\" sizes=\"auto, (max-width: 767px) 89vw, (max-width: 1000px) 54vw, (max-width: 1071px) 543px, 580px\" \/><figcaption>Test Pipeline <sup>[1]<\/sup><\/figcaption><\/figure>\n\n\n\n<p><strong>References:<\/strong><\/p>\n\n\n\n<p>[1] &#8220;WALT: Watch And Learn 2D Amodal Representation using Time-lapse Imagery&#8221;, N. Dinesh Reddy, Robert Tamburo, and Srinivasa Narasimha IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2022.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Datasets As mentioned earlier, there is no well-defined work zone targeted data available and gaining access to work-zone relevant data to train the models has been the main focus of our efforts. This can be achieved in two ways: Approach 1: Exploring existing datasets and models trained on them with at least a sub set &hellip; <\/p>\n<p class=\"link-more\"><a href=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/experimental-setup\/\" class=\"more-link\">Continue reading<span class=\"screen-reader-text\"> &#8220;Experimental Setup&#8221;<\/span><\/a><\/p>\n","protected":false},"author":68,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"open","template":"","meta":{"footnotes":""},"class_list":["post-2","page","type-page","status-publish","hentry"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v27.4 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>Experimental Setup - Platform Pittsburgh, Smart Cities<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/experimental-setup\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Experimental Setup - Platform Pittsburgh, Smart Cities\" \/>\n<meta property=\"og:description\" content=\"Datasets As mentioned earlier, there is no well-defined work zone targeted data available and gaining access to work-zone relevant data to train the models has been the main focus of our efforts. This can be achieved in two ways: Approach 1: Exploring existing datasets and models trained on them with at least a sub set &hellip; Continue reading &quot;Experimental Setup&quot;\" \/>\n<meta property=\"og:url\" content=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/experimental-setup\/\" \/>\n<meta property=\"og:site_name\" content=\"Platform Pittsburgh, Smart Cities\" \/>\n<meta property=\"article:modified_time\" content=\"2022-12-20T16:10:53+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.22.16-PM.png\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data1\" content=\"6 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/experimental-setup\\\/\",\"url\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/experimental-setup\\\/\",\"name\":\"Experimental Setup - Platform Pittsburgh, Smart Cities\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/experimental-setup\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/experimental-setup\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/wp-content\\\/uploads\\\/sites\\\/64\\\/2022\\\/04\\\/Screen-Shot-2022-04-26-at-7.22.16-PM.png\",\"datePublished\":\"2022-01-07T14:12:52+00:00\",\"dateModified\":\"2022-12-20T16:10:53+00:00\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/experimental-setup\\\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/experimental-setup\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/experimental-setup\\\/#primaryimage\",\"url\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/wp-content\\\/uploads\\\/sites\\\/64\\\/2022\\\/04\\\/Screen-Shot-2022-04-26-at-7.22.16-PM.png\",\"contentUrl\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/wp-content\\\/uploads\\\/sites\\\/64\\\/2022\\\/04\\\/Screen-Shot-2022-04-26-at-7.22.16-PM.png\",\"width\":485,\"height\":556},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/experimental-setup\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Experimental Setup\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/#website\",\"url\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/\",\"name\":\"Platform Pittsburgh, Smart Cities\",\"description\":\"Work-Zone Detection in Surveillance Cameras\",\"publisher\":{\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/#organization\",\"name\":\"Platform Pittsburgh, Smart Cities\",\"url\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/wp-content\\\/uploads\\\/sites\\\/64\\\/2022\\\/12\\\/cropped-image-removebg-preview.png\",\"contentUrl\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/wp-content\\\/uploads\\\/sites\\\/64\\\/2022\\\/12\\\/cropped-image-removebg-preview.png\",\"width\":250,\"height\":250,\"caption\":\"Platform Pittsburgh, Smart Cities\"},\"image\":{\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2022team9\\\/#\\\/schema\\\/logo\\\/image\\\/\"}}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Experimental Setup - Platform Pittsburgh, Smart Cities","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/experimental-setup\/","og_locale":"en_US","og_type":"article","og_title":"Experimental Setup - Platform Pittsburgh, Smart Cities","og_description":"Datasets As mentioned earlier, there is no well-defined work zone targeted data available and gaining access to work-zone relevant data to train the models has been the main focus of our efforts. This can be achieved in two ways: Approach 1: Exploring existing datasets and models trained on them with at least a sub set &hellip; Continue reading \"Experimental Setup\"","og_url":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/experimental-setup\/","og_site_name":"Platform Pittsburgh, Smart Cities","article_modified_time":"2022-12-20T16:10:53+00:00","og_image":[{"url":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.22.16-PM.png","type":"","width":"","height":""}],"twitter_card":"summary_large_image","twitter_misc":{"Est. reading time":"6 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/experimental-setup\/","url":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/experimental-setup\/","name":"Experimental Setup - Platform Pittsburgh, Smart Cities","isPartOf":{"@id":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/#website"},"primaryImageOfPage":{"@id":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/experimental-setup\/#primaryimage"},"image":{"@id":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/experimental-setup\/#primaryimage"},"thumbnailUrl":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.22.16-PM.png","datePublished":"2022-01-07T14:12:52+00:00","dateModified":"2022-12-20T16:10:53+00:00","breadcrumb":{"@id":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/experimental-setup\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/experimental-setup\/"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/experimental-setup\/#primaryimage","url":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.22.16-PM.png","contentUrl":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/04\/Screen-Shot-2022-04-26-at-7.22.16-PM.png","width":485,"height":556},{"@type":"BreadcrumbList","@id":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/experimental-setup\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/"},{"@type":"ListItem","position":2,"name":"Experimental Setup"}]},{"@type":"WebSite","@id":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/#website","url":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/","name":"Platform Pittsburgh, Smart Cities","description":"Work-Zone Detection in Surveillance Cameras","publisher":{"@id":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/#organization","name":"Platform Pittsburgh, Smart Cities","url":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/#\/schema\/logo\/image\/","url":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/cropped-image-removebg-preview.png","contentUrl":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-content\/uploads\/sites\/64\/2022\/12\/cropped-image-removebg-preview.png","width":250,"height":250,"caption":"Platform Pittsburgh, Smart Cities"},"image":{"@id":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/#\/schema\/logo\/image\/"}}]}},"_links":{"self":[{"href":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-json\/wp\/v2\/pages\/2","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-json\/wp\/v2\/users\/68"}],"replies":[{"embeddable":true,"href":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-json\/wp\/v2\/comments?post=2"}],"version-history":[{"count":47,"href":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-json\/wp\/v2\/pages\/2\/revisions"}],"predecessor-version":[{"id":326,"href":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-json\/wp\/v2\/pages\/2\/revisions\/326"}],"wp:attachment":[{"href":"https:\/\/mscvprojects.ri.cmu.edu\/2022team9\/wp-json\/wp\/v2\/media?parent=2"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}