{"id":49563,"date":"2021-08-17T11:30:47","date_gmt":"2021-08-17T03:30:47","guid":{"rendered":"https:\/\/www.seeedstudio.com\/blog\/?p=49563"},"modified":"2022-02-24T09:26:51","modified_gmt":"2022-02-24T01:26:51","slug":"multi-stage-inference-with-edge-impulse-tensorflow-lite-reterminal-raspberry-pi-cm4-inside","status":"publish","type":"post","link":"https:\/\/www.seeedstudio.com\/blog\/2021\/08\/17\/multi-stage-inference-with-edge-impulse-tensorflow-lite-reterminal-raspberry-pi-cm4-inside\/","title":{"rendered":"Multi-stage inference with Edge Impulse\/Tensorflow Lite &#8211; reTerminal (Raspberry Pi CM4 inside)"},"content":{"rendered":"\n<p>In comments to our articles and videos we get occasional questions about such tasks as face recognition, license plate recognition, emotion\/gender recognition and etc. These are all cases when you want to utilize multi-stage inference. Multi-stage inference in Computer Vision most of the time involves a combination of object detection and image classification in a single multi-model pipeline.<\/p>\n\n\n\n<figure class=\"wp-block-embed is-type-video is-provider-youtube wp-block-embed-youtube wp-embed-aspect-16-9 wp-has-aspect-ratio\"><div class=\"wp-block-embed__wrapper\">\n<iframe title=\"Multi-stage inference with Edge Impulse\/Tensorflow Lite - Raspberry Pi 4 Compute Module\" width=\"640\" height=\"360\" src=\"https:\/\/www.youtube.com\/embed\/ELCO_gJVDyI?feature=oembed\" frameborder=\"0\" allow=\"accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share\" referrerpolicy=\"strict-origin-when-cross-origin\" allowfullscreen><\/iframe>\n<\/div><\/figure>\n\n\n\n<p>What is the main benefit of using multi-stage inference?<br>Object detection networks in general, and the ones used on embedded devices in particular, are not very good at distinguishing between multiple similar classes of objects. <\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img fetchpriority=\"high\" decoding=\"async\" width=\"888\" height=\"449\" src=\"https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-20.png\" alt=\"\" class=\"wp-image-49628\" srcset=\"https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-20.png 888w, https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-20-300x152.png 300w, https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-20-768x388.png 768w\" sizes=\"(max-width: 888px) 100vw, 888px\" \/><\/figure>\n\n\n\n<p>So, what we see very often is using a detection networks to detect a large class of objects, for example faces or cars or dogs and then cropping and re-scaling the result of detection to be fed to image classification network, which gives a much more nuanced output, for example, the emotion on the face, car model or dog breed.<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" width=\"1030\" height=\"579\" src=\"https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-17-1030x579.png\" alt=\"\" class=\"wp-image-49625\" srcset=\"https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-17-1030x579.png 1030w, https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-17-300x169.png 300w, https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-17-768x432.png 768w, https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-17-1536x864.png 1536w, https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-17-1024x576.png 1024w, https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-17.png 1920w\" sizes=\"(max-width: 1030px) 100vw, 1030px\" \/><\/figure>\n\n\n\n<p>Another popular use case for multi-stage inference is OCR (or optical character recognition) &#8211; instead of performing detection of every character in the whole image, which is resource-intensive and prone to errors, most often the text is first detection and then recognition performed on pieces of text cropped out of the large image.<\/p>\n\n\n\n<p>For this article, let\u2019s take car model recognition as an example. In Edge Impulse train a car detector &#8211; this should be fairly easy, since we are using transfer learning and the base model already was trained to detect cars, so it contains the necessary feature maps. <\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" width=\"1030\" height=\"535\" src=\"https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-18-1030x535.png\" alt=\"\" class=\"wp-image-49626\" srcset=\"https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-18-1030x535.png 1030w, https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-18-300x156.png 300w, https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-18-768x399.png 768w, https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-18-1536x797.png 1536w, https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-18-1024x531.png 1024w, https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-18.png 1734w\" sizes=\"(max-width: 1030px) 100vw, 1030px\" \/><\/figure>\n\n\n\n<p>For image recognition model training, we\u2019re going to use a subset of Stanford Cars dataset, with 6 car model classes &#8211; I picked 6 classes of cars, that I think I\u2019ll likely to encounter while walking outside in Shenzhen, China. <\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1030\" height=\"525\" src=\"https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-19-1030x525.png\" alt=\"\" class=\"wp-image-49627\" srcset=\"https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-19-1030x525.png 1030w, https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-19-300x153.png 300w, https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-19-768x391.png 768w, https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-19-1536x782.png 1536w, https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-19-1024x521.png 1024w, https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/image-19.png 1677w\" sizes=\"(max-width: 1030px) 100vw, 1030px\" \/><\/figure>\n\n\n\n<p>After training the models and optionally, checking the detection and recognition accuracy on pictures from the internet, let\u2019s deploy them to an embedded device. I have used reTerminal, a Raspberry Pi Compute Module 4 based development board with a touchscreen in a sturdy plastic case &#8211; it comes in handy while on field trips like this one. <\/p>\n\n\n\n<figure class=\"wp-block-embed is-type-rich is-provider-twitter wp-block-embed-twitter\"><div class=\"wp-block-embed__wrapper\">\n<blockquote class=\"twitter-tweet\" data-width=\"550\" data-dnt=\"true\"><p lang=\"en\" dir=\"ltr\">reTerminal from <a href=\"https:\/\/twitter.com\/seeedstudio?ref_src=twsrc%5Etfw\">@seeedstudio<\/a>, a <a href=\"https:\/\/twitter.com\/Raspberry_Pi?ref_src=twsrc%5Etfw\">@Raspberry_Pi<\/a> CM 4  based development board with a touchscreen in a sturdy plastic case &#8211; it comes in handy while on field trips like this one<br>Multi-stage inference with <a href=\"https:\/\/twitter.com\/EdgeImpulse?ref_src=twsrc%5Etfw\">@EdgeImpulse<\/a> Linux Python SDK Object detection -&gt; Fine-grained classification <a href=\"https:\/\/t.co\/jF7PNGeDLu\">pic.twitter.com\/jF7PNGeDLu<\/a><\/p>&mdash; hardware-ai (@HardwareAi) <a href=\"https:\/\/twitter.com\/HardwareAi\/status\/1415325930352873477?ref_src=twsrc%5Etfw\">July 14, 2021<\/a><\/blockquote><script async src=\"https:\/\/platform.twitter.com\/widgets.js\" charset=\"utf-8\"><\/script>\n<\/div><\/figure>\n\n\n\n<p>Raspberry Pi Compute Module 4 has the same CPU as Raspbery Pi 4 development board, but has an option to include onboard eMMC memory &#8211; the module in reTerminal has 4 GB or RAM and 32 GB of eMMC.<\/p>\n\n\n\n<p>First of all you\u2019ll need to download both detection and classification models with the help of edge-impulse-linux-runner. To install edge-impulse-runner on reTerminal (or Raspberry Pi) consult the <a aria-label=\"undefined (opens in a new tab)\" href=\"https:\/\/docs.edgeimpulse.com\/docs\/raspberry-pi-4\" target=\"_blank\" rel=\"noreferrer noopener\">official documentation<\/a>. After it is installed run the following commands to download models:<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>edge-impulse-linux-runner --download car_detector.eim\nedge-impulse-linux-runner --clean\nedge-impulse-linux-runner --download car_classifier.eim<\/code><\/pre>\n\n\n\n<p>The multi-stage inference script is very similar to image classification\/object detection scripts, except as you might have guessed it has both models &#8211; you can find the code in <a href=\"https:\/\/github.com\/Seeed-Studio\/Seeed_Python_MachineLearning\/tree\/main\/examples\/edge_impulse\/multi_stage_inference_vehicle_type\" target=\"_blank\" aria-label=\"undefined (opens in a new tab)\" rel=\"noreferrer noopener\">this Github repository<\/a>. Remember to install all the dependencies before running the code:<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>sudo apt-get install libatlas-base-dev libportaudio0 libportaudio2 libportaudiocpp0 portaudio19-dev\npip3 install install edge_impulse_linux -i https:\/\/pypi.python.org\/simple\npython3 multi_stage.py car_detector.eim car_classifier.eim <\/code><\/pre>\n\n\n\n<p>Alternatively, you can do multi-stage inference using another inference framework of your choice. For this have a look at <a aria-label=\"undefined (opens in a new tab)\" href=\"https:\/\/github.com\/Seeed-Studio\/Seeed_Python_MachineLearning\/blob\/main\/jupyter_notebooks\/aXeleRate_multi_stage.ipynb\" target=\"_blank\" rel=\"noreferrer noopener\">Jupyter Notebook<\/a> I prepared, which uses <a aria-label=\"undefined (opens in a new tab)\" href=\"https:\/\/github.com\/AIWintermuteAI\/aXeleRate\" target=\"_blank\" rel=\"noreferrer noopener\">aXeleRate, a Keras-based framework for AI on the edge<\/a>. There I trained small MobileNet v1 alpha 0.25 YOLOv3 to detect vehicles and slightly larger MobileNet v1 alpha 0.5 for classifying all cars from Standford cars dataset. Jupyter Notebook includes training examples &#8211; if you\u2019d like to run them on a Raspberry Pi or reTerminal, download the <a aria-label=\"undefined (opens in a new tab)\" href=\"https:\/\/github.com\/Seeed-Studio\/Seeed_Python_MachineLearning\/tree\/main\/examples\/tensorflow_lite\/multi_stage_inference_vehicle_type\" target=\"_blank\" rel=\"noreferrer noopener\">code from here<\/a> and run <\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>python3 multi_stage_file.py --first_stage yolo_best_recall.tflite --second_stage classifier_best_accuracy.tflite --labels labels.txt  --file ..\/..\/sample_files\/cars.mp4 <\/code><\/pre>\n\n\n\n<p>for running inference on video file and <\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>python3 multi_stage_stream.py --first_stage yolo_best_recall.tflite --second_stage classifier_best_accuracy.tflite --labels labels.txt <\/code><\/pre>\n\n\n\n<p>To run inference on video stream from web camera.<\/p>\n\n\n\n<p>In conclusion, do we really need multi-stage inference and could object detector do both detection and fine-grained classification? There are some models, that can perform detection and large class-size classification pretty well, such as YOLO9000, but these are not suitable for running on embedded devices with constrained resources. So, for now, multi-stage inference is the best technique we have in our Computer Vision arsenals for this type of applications.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>In comments to our articles and videos we get occasional questions about such tasks as<\/p>\n","protected":false},"author":3505,"featured_media":49634,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_lmt_disableupdate":"","_lmt_disable":"","_price":"","_stock":"","_tribe_ticket_header":"","_tribe_default_ticket_provider":"","_tribe_ticket_capacity":"0","_ticket_start_date":"","_ticket_end_date":"","_tribe_ticket_show_description":"","_tribe_ticket_show_not_going":false,"_tribe_ticket_use_global_stock":"","_tribe_ticket_global_stock_level":"","_global_stock_mode":"","_global_stock_cap":"","_tribe_rsvp_for_event":"","_tribe_ticket_going_count":"","_tribe_ticket_not_going_count":"","_tribe_tickets_list":"[]","_tribe_ticket_has_attendee_info_fields":false,"iawp_total_views":0,"footnotes":""},"categories":[1],"tags":[3254,1302,3494,3820],"class_list":["post-49563","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-news","tag-computer-vision","tag-edge-computing","tag-raspberry-pi-compute-module-4","tag-reterminal"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v24.0 - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>Multi-stage inference with Edge Impulse\/Tensorflow Lite - reTerminal (Raspberry Pi CM4 inside) - Latest News from Seeed Studio<\/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:\/\/www.seeedstudio.com\/blog\/2021\/08\/17\/multi-stage-inference-with-edge-impulse-tensorflow-lite-reterminal-raspberry-pi-cm4-inside\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Multi-stage inference with Edge Impulse\/Tensorflow Lite - reTerminal (Raspberry Pi CM4 inside) - Latest News from Seeed Studio\" \/>\n<meta property=\"og:description\" content=\"In comments to our articles and videos we get occasional questions about such tasks as\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.seeedstudio.com\/blog\/2021\/08\/17\/multi-stage-inference-with-edge-impulse-tensorflow-lite-reterminal-raspberry-pi-cm4-inside\/\" \/>\n<meta property=\"og:site_name\" content=\"Latest News from Seeed Studio\" \/>\n<meta property=\"article:published_time\" content=\"2021-08-17T03:30:47+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2022-02-24T01:26:51+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/Presentation1-e1629171194545.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"1136\" \/>\n\t<meta property=\"og:image:height\" content=\"900\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"author\" content=\"Dmitry Maslov\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Dmitry Maslov\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"4 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/www.seeedstudio.com\/blog\/2021\/08\/17\/multi-stage-inference-with-edge-impulse-tensorflow-lite-reterminal-raspberry-pi-cm4-inside\/\",\"url\":\"https:\/\/www.seeedstudio.com\/blog\/2021\/08\/17\/multi-stage-inference-with-edge-impulse-tensorflow-lite-reterminal-raspberry-pi-cm4-inside\/\",\"name\":\"Multi-stage inference with Edge Impulse\/Tensorflow Lite - reTerminal (Raspberry Pi CM4 inside) - Latest News from Seeed Studio\",\"isPartOf\":{\"@id\":\"https:\/\/www.seeedstudio.com\/blog\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\/\/www.seeedstudio.com\/blog\/2021\/08\/17\/multi-stage-inference-with-edge-impulse-tensorflow-lite-reterminal-raspberry-pi-cm4-inside\/#primaryimage\"},\"image\":{\"@id\":\"https:\/\/www.seeedstudio.com\/blog\/2021\/08\/17\/multi-stage-inference-with-edge-impulse-tensorflow-lite-reterminal-raspberry-pi-cm4-inside\/#primaryimage\"},\"thumbnailUrl\":\"https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/Presentation1-e1629171194545.jpg\",\"datePublished\":\"2021-08-17T03:30:47+00:00\",\"dateModified\":\"2022-02-24T01:26:51+00:00\",\"author\":{\"@id\":\"https:\/\/www.seeedstudio.com\/blog\/#\/schema\/person\/be44021cef50367de429a4d5f613ed2f\"},\"breadcrumb\":{\"@id\":\"https:\/\/www.seeedstudio.com\/blog\/2021\/08\/17\/multi-stage-inference-with-edge-impulse-tensorflow-lite-reterminal-raspberry-pi-cm4-inside\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/www.seeedstudio.com\/blog\/2021\/08\/17\/multi-stage-inference-with-edge-impulse-tensorflow-lite-reterminal-raspberry-pi-cm4-inside\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/www.seeedstudio.com\/blog\/2021\/08\/17\/multi-stage-inference-with-edge-impulse-tensorflow-lite-reterminal-raspberry-pi-cm4-inside\/#primaryimage\",\"url\":\"https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/Presentation1-e1629171194545.jpg\",\"contentUrl\":\"https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/Presentation1-e1629171194545.jpg\",\"width\":1136,\"height\":900},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/www.seeedstudio.com\/blog\/2021\/08\/17\/multi-stage-inference-with-edge-impulse-tensorflow-lite-reterminal-raspberry-pi-cm4-inside\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/www.seeedstudio.com\/blog\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Multi-stage inference with Edge Impulse\/Tensorflow Lite &#8211; reTerminal (Raspberry Pi CM4 inside)\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/www.seeedstudio.com\/blog\/#website\",\"url\":\"https:\/\/www.seeedstudio.com\/blog\/\",\"name\":\"Latest News from Seeed Studio\",\"description\":\"Emerging IoT, AI and Autonomous Applications on the Edge\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/www.seeedstudio.com\/blog\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/www.seeedstudio.com\/blog\/#\/schema\/person\/be44021cef50367de429a4d5f613ed2f\",\"name\":\"Dmitry Maslov\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/www.seeedstudio.com\/blog\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/secure.gravatar.com\/avatar\/b60714970fdc7dfa4a5d9915477bdd24?s=96&r=g\",\"contentUrl\":\"https:\/\/secure.gravatar.com\/avatar\/b60714970fdc7dfa4a5d9915477bdd24?s=96&r=g\",\"caption\":\"Dmitry Maslov\"},\"url\":\"https:\/\/www.seeedstudio.com\/blog\/author\/dmitry-maslov\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Multi-stage inference with Edge Impulse\/Tensorflow Lite - reTerminal (Raspberry Pi CM4 inside) - Latest News from Seeed Studio","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:\/\/www.seeedstudio.com\/blog\/2021\/08\/17\/multi-stage-inference-with-edge-impulse-tensorflow-lite-reterminal-raspberry-pi-cm4-inside\/","og_locale":"en_US","og_type":"article","og_title":"Multi-stage inference with Edge Impulse\/Tensorflow Lite - reTerminal (Raspberry Pi CM4 inside) - Latest News from Seeed Studio","og_description":"In comments to our articles and videos we get occasional questions about such tasks as","og_url":"https:\/\/www.seeedstudio.com\/blog\/2021\/08\/17\/multi-stage-inference-with-edge-impulse-tensorflow-lite-reterminal-raspberry-pi-cm4-inside\/","og_site_name":"Latest News from Seeed Studio","article_published_time":"2021-08-17T03:30:47+00:00","article_modified_time":"2022-02-24T01:26:51+00:00","og_image":[{"width":1136,"height":900,"url":"https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/Presentation1-e1629171194545.jpg","type":"image\/jpeg"}],"author":"Dmitry Maslov","twitter_card":"summary_large_image","twitter_misc":{"Written by":"Dmitry Maslov","Est. reading time":"4 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/www.seeedstudio.com\/blog\/2021\/08\/17\/multi-stage-inference-with-edge-impulse-tensorflow-lite-reterminal-raspberry-pi-cm4-inside\/","url":"https:\/\/www.seeedstudio.com\/blog\/2021\/08\/17\/multi-stage-inference-with-edge-impulse-tensorflow-lite-reterminal-raspberry-pi-cm4-inside\/","name":"Multi-stage inference with Edge Impulse\/Tensorflow Lite - reTerminal (Raspberry Pi CM4 inside) - Latest News from Seeed Studio","isPartOf":{"@id":"https:\/\/www.seeedstudio.com\/blog\/#website"},"primaryImageOfPage":{"@id":"https:\/\/www.seeedstudio.com\/blog\/2021\/08\/17\/multi-stage-inference-with-edge-impulse-tensorflow-lite-reterminal-raspberry-pi-cm4-inside\/#primaryimage"},"image":{"@id":"https:\/\/www.seeedstudio.com\/blog\/2021\/08\/17\/multi-stage-inference-with-edge-impulse-tensorflow-lite-reterminal-raspberry-pi-cm4-inside\/#primaryimage"},"thumbnailUrl":"https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/Presentation1-e1629171194545.jpg","datePublished":"2021-08-17T03:30:47+00:00","dateModified":"2022-02-24T01:26:51+00:00","author":{"@id":"https:\/\/www.seeedstudio.com\/blog\/#\/schema\/person\/be44021cef50367de429a4d5f613ed2f"},"breadcrumb":{"@id":"https:\/\/www.seeedstudio.com\/blog\/2021\/08\/17\/multi-stage-inference-with-edge-impulse-tensorflow-lite-reterminal-raspberry-pi-cm4-inside\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.seeedstudio.com\/blog\/2021\/08\/17\/multi-stage-inference-with-edge-impulse-tensorflow-lite-reterminal-raspberry-pi-cm4-inside\/"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/www.seeedstudio.com\/blog\/2021\/08\/17\/multi-stage-inference-with-edge-impulse-tensorflow-lite-reterminal-raspberry-pi-cm4-inside\/#primaryimage","url":"https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/Presentation1-e1629171194545.jpg","contentUrl":"https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/Presentation1-e1629171194545.jpg","width":1136,"height":900},{"@type":"BreadcrumbList","@id":"https:\/\/www.seeedstudio.com\/blog\/2021\/08\/17\/multi-stage-inference-with-edge-impulse-tensorflow-lite-reterminal-raspberry-pi-cm4-inside\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/www.seeedstudio.com\/blog\/"},{"@type":"ListItem","position":2,"name":"Multi-stage inference with Edge Impulse\/Tensorflow Lite &#8211; reTerminal (Raspberry Pi CM4 inside)"}]},{"@type":"WebSite","@id":"https:\/\/www.seeedstudio.com\/blog\/#website","url":"https:\/\/www.seeedstudio.com\/blog\/","name":"Latest News from Seeed Studio","description":"Emerging IoT, AI and Autonomous Applications on the Edge","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.seeedstudio.com\/blog\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Person","@id":"https:\/\/www.seeedstudio.com\/blog\/#\/schema\/person\/be44021cef50367de429a4d5f613ed2f","name":"Dmitry Maslov","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/www.seeedstudio.com\/blog\/#\/schema\/person\/image\/","url":"https:\/\/secure.gravatar.com\/avatar\/b60714970fdc7dfa4a5d9915477bdd24?s=96&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/b60714970fdc7dfa4a5d9915477bdd24?s=96&r=g","caption":"Dmitry Maslov"},"url":"https:\/\/www.seeedstudio.com\/blog\/author\/dmitry-maslov\/"}]}},"modified_by":"Amanda Sun","views":4537,"featured_image_urls":{"full":["https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/Presentation1-e1629171194545.jpg",1136,900,false],"thumbnail":["https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/Presentation1-e1629171194545-80x80.jpg",80,80,true],"medium":["https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/Presentation1-e1629171194545-300x238.jpg",300,238,true],"medium_large":["https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/Presentation1-e1629171194545-768x608.jpg",640,507,true],"large":["https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/Presentation1-e1629171194545-1030x816.jpg",640,507,true],"1536x1536":["https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/Presentation1-1536x864.jpg",1536,864,true],"2048x2048":["https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/Presentation1-e1629171194545.jpg",1136,900,false],"visody_icon":["https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/Presentation1-e1629171194545.jpg",32,25,false],"magazine-7-slider-full":["https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/Presentation1-1536x900.jpg",1536,900,true],"magazine-7-slider-center":["https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/Presentation1-e1629171194545-936x897.jpg",936,897,true],"magazine-7-featured":["https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/Presentation1-e1629171194545-1024x811.jpg",1024,811,true],"magazine-7-medium":["https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/Presentation1-e1629171194545-720x380.jpg",720,380,true],"magazine-7-medium-square":["https:\/\/www.seeedstudio.com\/blog\/wp-content\/uploads\/2021\/08\/Presentation1-e1629171194545-675x450.jpg",675,450,true]},"author_info":{"display_name":"Dmitry Maslov","author_link":"https:\/\/www.seeedstudio.com\/blog\/author\/dmitry-maslov\/"},"category_info":"<a href=\"https:\/\/www.seeedstudio.com\/blog\/category\/news\/\" rel=\"category tag\">News<\/a>","tag_info":"News","comment_count":"0","_links":{"self":[{"href":"https:\/\/www.seeedstudio.com\/blog\/wp-json\/wp\/v2\/posts\/49563","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.seeedstudio.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.seeedstudio.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.seeedstudio.com\/blog\/wp-json\/wp\/v2\/users\/3505"}],"replies":[{"embeddable":true,"href":"https:\/\/www.seeedstudio.com\/blog\/wp-json\/wp\/v2\/comments?post=49563"}],"version-history":[{"count":8,"href":"https:\/\/www.seeedstudio.com\/blog\/wp-json\/wp\/v2\/posts\/49563\/revisions"}],"predecessor-version":[{"id":60880,"href":"https:\/\/www.seeedstudio.com\/blog\/wp-json\/wp\/v2\/posts\/49563\/revisions\/60880"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.seeedstudio.com\/blog\/wp-json\/wp\/v2\/media\/49634"}],"wp:attachment":[{"href":"https:\/\/www.seeedstudio.com\/blog\/wp-json\/wp\/v2\/media?parent=49563"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.seeedstudio.com\/blog\/wp-json\/wp\/v2\/categories?post=49563"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.seeedstudio.com\/blog\/wp-json\/wp\/v2\/tags?post=49563"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}