{"id":984,"date":"2024-02-18T05:19:11","date_gmt":"2024-02-18T05:19:11","guid":{"rendered":"https:\/\/musefind.com\/?p=984"},"modified":"2024-02-18T07:13:03","modified_gmt":"2024-02-18T07:13:03","slug":"what-we-learned-from-3-best-image-analysis-api","status":"publish","type":"post","link":"https:\/\/musefind.com\/what-we-learned-from-3-best-image-analysis-api\/","title":{"rendered":"We compared the 3 best image analysis API\u2019s \u2014 here\u2019s what we learned"},"content":{"rendered":"\n<div class=\"wp-block-rank-math-toc-block\" id=\"rank-math-toc\"><div><strong>In This Article:<\/strong><\/div><nav><ul><li class=\"\"><a href=\"#clarifai\">Clarifai<\/a><\/li><li class=\"\"><a href=\"#cloud-vision-api\">Cloud Vision API<\/a><\/li><li class=\"\"><a href=\"#amazon-rekognition\">Amazon Rekognition<\/a><\/li><li class=\"\"><a href=\"#conclusion\">Conclusion<\/a><\/li><\/ul><\/nav><\/div>\n\n\n\n<p id=\"9eab\">One of the flagship applications of modern machine learning has been working with images \u2014 training computers to analyze, classify, and alter different types of pictures. Last year, Google\u2019s DeepDream software made waves by creating a series of terrifying, nightmare-inducing images:<\/p>\n\n\n\n<figure class=\"wp-block-image size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"720\" height=\"405\" src=\"https:\/\/musefind.com\/wp-content\/uploads\/2024\/02\/image1.jpg\" alt=\"image1\" class=\"wp-image-988\" srcset=\"https:\/\/musefind.com\/wp-content\/uploads\/2024\/02\/image1.jpg 720w, https:\/\/musefind.com\/wp-content\/uploads\/2024\/02\/image1-300x169.jpg 300w\" sizes=\"(max-width: 720px) 100vw, 720px\" \/><\/figure>\n\n\n\n<p id=\"4903\">Thanks, Google.<\/p>\n\n\n\n<p id=\"560c\">While image and video classification (that is, telling us what is&nbsp;<em>depicted in&nbsp;<\/em>the chosen media) is not so cutting edge anymore, this is actually good news. A number of cheap services have sprung up to make image classification quite accessible.<\/p>\n\n\n\n<p id=\"eebf\">Here at MuseFind we\u2019ve been looking for ways to help both brands and social media influencers produce better content \u2014 and part of that effort is figuring out the correlations between the most successful posts and their content.<\/p>\n\n\n\n<p id=\"3de4\">Here\u2019s our example picture, from one of MuseFind\u2019s <a href=\"https:\/\/musefind.com\/influencer-marketing-latest-trends-best-practices-2018-report\/\">influencer marketing<\/a> campaigns with John Varvatos. Courtesy of&nbsp;<a href=\"https:\/\/www.instagram.com\/rusty.blade\/\" target=\"_blank\" rel=\"noreferrer noopener\">@rusty.blade<\/a>:<\/p>\n\n\n\n<figure class=\"wp-block-image size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"720\" height=\"750\" src=\"https:\/\/musefind.com\/wp-content\/uploads\/2024\/02\/image2.jpg\" alt=\"image2\" class=\"wp-image-989\" srcset=\"https:\/\/musefind.com\/wp-content\/uploads\/2024\/02\/image2.jpg 720w, https:\/\/musefind.com\/wp-content\/uploads\/2024\/02\/image2-288x300.jpg 288w\" sizes=\"(max-width: 720px) 100vw, 720px\" \/><\/figure>\n\n\n\n<p id=\"03aa\">Below are the top three image\/video analysis services we considered: Clarifai, Google Cloud Vision, and Amazon Rekognition. Let\u2019s see what they can tell us about this image.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"clarifai\">Clarifai<\/h2>\n\n\n\n<p id=\"f862\">Clarifai, unlike any other API\u2019s on this list, has the added bonus of video analysis. View a demo below:<\/p>\n\n\n\n<p id=\"2508\">For videos, Clarifai offers scene recognition. For images, it can also do sentiment analysis, text recognition, logo detection, and face detection, as well as a more robust version of Resemble\u2019s image attribute detection: brightness, colour, dominant colour.<\/p>\n\n\n\n<p id=\"66a8\">Here\u2019s what it shows about our image:<\/p>\n\n\n\n<figure class=\"wp-block-image size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"476\" height=\"511\" src=\"https:\/\/musefind.com\/wp-content\/uploads\/2024\/02\/Resembles-image-attribute-detection.jpg\" alt=\"Resemble\u2019s image attribute detection\" class=\"wp-image-991\" srcset=\"https:\/\/musefind.com\/wp-content\/uploads\/2024\/02\/Resembles-image-attribute-detection.jpg 476w, https:\/\/musefind.com\/wp-content\/uploads\/2024\/02\/Resembles-image-attribute-detection-279x300.jpg 279w\" sizes=\"(max-width: 476px) 100vw, 476px\" \/><\/figure>\n\n\n\n<p id=\"6590\">It was able to recognize some key concepts (the first five are fairly accurate), but also added some irrelevant tags. No scarf, no child, no woman.<\/p>\n\n\n\n<p id=\"435c\">Still, if we took the top five tags and used that to automatically categorize our image, we\u2019d be in a good place.<\/p>\n\n\n\n<p id=\"7134\"><strong>Price: 5,000 free operations, 10,000 free inputs per month- $1.20\/$.80 per 1000 after that<\/strong><\/p>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"cloud-vision-api\">Cloud Vision API<\/h2>\n\n\n\n<p id=\"7a4c\">Cloud Vision by Google includes many of the key features of Clarifai (sentiment analysis, text recognition, logo detection, facial analysis) and adds a couple of bonuses: landmark detection and a simple REST API.<\/p>\n\n\n\n<p><a href=\"https:\/\/medium.com\/google-cloud\/google-cloud-vision-landmark-detection-vacation-photos-41392d4b5765\" target=\"_blank\" rel=\"noopener\">https:\/\/medium.com\/google-cloud\/google-cloud-vision-landmark-detection-vacation-photos-41392d4b5765<\/a><\/p>\n\n\n\n<p id=\"ec17\">Unlike Clarifai, you can\u2019t create your own models to test against \u2014 but you also have access to an API backed by Google, one they seem quite dedicated to constantly improving.<\/p>\n\n\n\n<p id=\"bce6\">For our image, it gave us the following labels:<\/p>\n\n\n\n<figure class=\"wp-block-image size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"720\" height=\"358\" src=\"https:\/\/musefind.com\/wp-content\/uploads\/2024\/02\/Cloud-Vision-was-accurate.jpg\" alt=\"Cloud Vision was accurate\" class=\"wp-image-986\" srcset=\"https:\/\/musefind.com\/wp-content\/uploads\/2024\/02\/Cloud-Vision-was-accurate.jpg 720w, https:\/\/musefind.com\/wp-content\/uploads\/2024\/02\/Cloud-Vision-was-accurate-300x149.jpg 300w\" sizes=\"(max-width: 720px) 100vw, 720px\" \/><\/figure>\n\n\n\n<p id=\"f83e\">Cloud Vision was accurate in all but one tag (Spring) and managed to identify not just the dog but the activity of dog walking.<\/p>\n\n\n\n<p id=\"5e7c\">However, it did fail to detect the face in the image. Here\u2019s a preview of its sentiment analysis with a different image:<\/p>\n\n\n\n<figure class=\"wp-block-image size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"720\" height=\"300\" src=\"https:\/\/musefind.com\/wp-content\/uploads\/2024\/02\/sentiment-analysis-with-a-different-image.jpg\" alt=\"sentiment analysis with a different image\" class=\"wp-image-985\" srcset=\"https:\/\/musefind.com\/wp-content\/uploads\/2024\/02\/sentiment-analysis-with-a-different-image.jpg 720w, https:\/\/musefind.com\/wp-content\/uploads\/2024\/02\/sentiment-analysis-with-a-different-image-300x125.jpg 300w\" sizes=\"(max-width: 720px) 100vw, 720px\" \/><\/figure>\n\n\n\n<p id=\"fa1b\"><strong>Price: Free up to 1000\/month, $1.50 per thousand beyond that<\/strong><\/p>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"amazon-rekognition\">Amazon Rekognition<\/h2>\n\n\n\n<p id=\"e419\">By now, the features will sound familiar: scene detection, sentiment analysis. Amazon boasts of a more robust suite of facial analysis tools, including facial recognition (not offered by Google or Clarifai) across images, and detailed information like image quality of face, beard recognition (yes), and facial comparison (how likely is it that two faces are the same person?).<\/p>\n\n\n\n<p id=\"bb26\">It also promises integration with AWS services like S3 and Lamba.<\/p>\n\n\n\n<p><a href=\"https:\/\/aws.amazon.com\/rekognition\/\" target=\"_blank\" rel=\"noopener\">https:\/\/aws.amazon.com\/rekognition\/<\/a><\/p>\n\n\n\n<p>Let\u2019s start with object and scene detection. Here\u2019s the results:<\/p>\n\n\n\n<figure class=\"wp-block-image size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"430\" height=\"568\" src=\"https:\/\/musefind.com\/wp-content\/uploads\/2024\/02\/object-and-scene-detection.jpg\" alt=\"object and scene detection\" class=\"wp-image-990\" srcset=\"https:\/\/musefind.com\/wp-content\/uploads\/2024\/02\/object-and-scene-detection.jpg 430w, https:\/\/musefind.com\/wp-content\/uploads\/2024\/02\/object-and-scene-detection-227x300.jpg 227w\" sizes=\"(max-width: 430px) 100vw, 430px\" \/><\/figure>\n\n\n\n<p id=\"49a4\">It nailed the dog aspect (bonus points for \u2018Adorable\u2019) but failed to recognize any other aspect of the scene.<\/p>\n\n\n\n<p id=\"3ce4\">However, it did pick up the face:<\/p>\n\n\n\n<figure class=\"wp-block-image size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"438\" height=\"628\" src=\"https:\/\/musefind.com\/wp-content\/uploads\/2024\/02\/face.jpg\" alt=\"face\" class=\"wp-image-987\" srcset=\"https:\/\/musefind.com\/wp-content\/uploads\/2024\/02\/face.jpg 438w, https:\/\/musefind.com\/wp-content\/uploads\/2024\/02\/face-209x300.jpg 209w\" sizes=\"(max-width: 438px) 100vw, 438px\" \/><\/figure>\n\n\n\n<p id=\"f8ce\">That killer beard detection. A bit of a mixed bag (no sunglasses, no mustache), but also reasonable accuracy.<\/p>\n\n\n\n<p id=\"f7a4\"><strong>Price: $1 per 1000 images processed (up to a million a month)<\/strong><\/p>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"conclusion\">Conclusion<\/h2>\n\n\n\n<p id=\"6e55\">We haven\u2019t settled on a service yet \u2014 we\u2019re going to continue experimenting and clarifying our goals over the coming weeks.<\/p>\n\n\n\n<p id=\"c0cb\">But based on these results, we already get an idea of the strengths and weaknesses of the different API\u2019s. Clarifai has the strongest concept modeling, Google the best scene detection and sentiment analysis, and Amazon the best facial analysis.<\/p>\n\n\n\n<p id=\"9560\">Let us know in the comments your thoughts on these API\u2019s, and whether you have any additional comments. And please like and share if this article was useful to you!<\/p>\n","protected":false},"excerpt":{"rendered":"<p>One of the flagship applications of modern machine learning has been working with images \u2014 training computers to analyze, classify, and alter different types of pictures. Last year, Google\u2019s DeepDream software made waves by creating a series of terrifying, nightmare-inducing images: Thanks, Google. While image and video classification (that is, telling us what is&nbsp;depicted in&nbsp;the &#8230; <a title=\"We compared the 3 best image analysis API\u2019s \u2014 here\u2019s what we learned\" class=\"read-more\" href=\"https:\/\/musefind.com\/what-we-learned-from-3-best-image-analysis-api\/\" aria-label=\"More on We compared the 3 best image analysis API\u2019s \u2014 here\u2019s what we learned\">Read more<\/a><\/p>\n","protected":false},"author":5,"featured_media":988,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[8],"tags":[],"_links":{"self":[{"href":"https:\/\/musefind.com\/wp-json\/wp\/v2\/posts\/984"}],"collection":[{"href":"https:\/\/musefind.com\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/musefind.com\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/musefind.com\/wp-json\/wp\/v2\/users\/5"}],"replies":[{"embeddable":true,"href":"https:\/\/musefind.com\/wp-json\/wp\/v2\/comments?post=984"}],"version-history":[{"count":4,"href":"https:\/\/musefind.com\/wp-json\/wp\/v2\/posts\/984\/revisions"}],"predecessor-version":[{"id":1034,"href":"https:\/\/musefind.com\/wp-json\/wp\/v2\/posts\/984\/revisions\/1034"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/musefind.com\/wp-json\/wp\/v2\/media\/988"}],"wp:attachment":[{"href":"https:\/\/musefind.com\/wp-json\/wp\/v2\/media?parent=984"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/musefind.com\/wp-json\/wp\/v2\/categories?post=984"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/musefind.com\/wp-json\/wp\/v2\/tags?post=984"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}