Digital Automated Identification System

<p>Digital automated identification system (DAISY) is an <a href="/facts/Automated_species_identification/eSPC0fSn">automated species identification</a> system optimised for the rapid screening of invertebrates (e.g. insects) by non-experts (e.g. <a href="/facts/Parataxonomy/f7omU8zv">parataxonomists</a>).
</p><p>It was developed by Dr. <a href="/facts/Mark_A._O%2527Neill/9IVqRp1c">Mark O'Neill</a> during the mid-1990s. Development was supported by funding from the <a href="/facts/Darwin_Initiative/JI3oHkdI">Darwin Initiative</a> in 1997 and <a href="/facts/BBSRC/vxtxKH5b">BBSRC</a>. The <a href="/facts/Intellectual_property_rights/PqHW9mVP">intellectual property rights</a> were acquired by O'Neill's company, Tumbling Dice Ltd, in February 2000 at the end of the grant funded <a href="http://www.darwininitiative.org.uk/project/6050/">Darwin Project</a>. The system underwent further development resulting in an producing an exemplar which is web accessible and which can cope in near real time with groups (e.g. <a href="/facts/Sphingidae/M0juzV0K">hawk moths</a>) which contain several hundred <a href="/facts/Taxa/caQC1e87">taxa</a>. On medium to high end PC server hardware (e.g. a <a href="/facts/Blade_server/Eo9gKdnU">blade server</a>) an identification is possible in under a second for a 300 taxon group. Parallelisation of the critical DAISY classifier codes (using either bespoke <a href="/facts/FPGA/M8LSGcAU">FPGA</a> technology or general purpose <a href="/facts/GPU/PTK1RQVp">GPU</a> programming technology such as <a href="/facts/CUDA/fvPpolot">CUDA</a>) will give an order of magnitude increase in performance. This means that DAISY can be deployed to make real time identifications within groups containing thousands of taxa (e.g. <a href="/facts/True_flies/ugzGBFKG">true flies</a>).
</p>
DAISY results for selected insect taxa<table><tbody><tr><th scope="col">taxon</th><th scope="col">image type</th><th scope="col">structure</th><th scope="col">training images</th><th scope="col">species</th><th scope="col">success(%)</th></tr><tr><td><a href="/facts/Belize/2xVuhCTt">Belize</a> <a href="/facts/Sphingidae/M0juzV0K">Sphingidae</a></td><td>RGB</td><td>wing</td><td>705</td><td>58</td><td>97</td></tr><tr><td><i><a href="/facts/Xylophanes/g4bJ7pch">Xylophanes</a></i> sp.</td><td>RGB</td><td>wing</td><td>543</td><td>30</td><td>99</td></tr><tr><td><a href="/facts/Parasitic_wasps/vLVmVHBs">Parasitic wasps</a></td><td>mono</td><td>wing</td><td>559</td><td>47</td><td>95</td></tr><tr><td>UK butterflies</td><td>RGB</td><td>wing</td><td>818</td><td>57</td><td>98</td></tr><tr><td>UK macro moths</td><td>RGB</td><td>wing</td><td>744</td><td>37</td><td>98</td></tr><tr><td>Caterpillars</td><td>RGB</td><td>head</td><td>91</td><td>7</td><td>93</td></tr><tr><td>Caterpillars</td><td>RGB</td><td>body</td><td>508</td><td>10</td><td>99</td></tr><tr><td>Soft fruit pests</td><td>RGB</td><td>body</td><td>2634</td><td>23</td><td>91</td></tr></tbody></table>
DAISY results for other image classification tasks<table><tbody><tr><th scope="col">description</th><th scope="col">image type</th><th scope="col">structure</th><th scope="col">training images</th><th scope="col">classes</th><th scope="col">success(%)</th></tr><tr><td>Food cans</td><td>RGB</td><td>label</td><td>31</td><td>5</td><td>100</td></tr><tr><td>Industrial objects</td><td>RGB</td><td>unposed object</td><td>155</td><td>14</td><td>100</td></tr><tr><td><a href="/facts/Foraminifera/8Yo9cPTI">Foraminifera</a> tests</td><td>RGB</td><td>unposed object</td><td>198</td><td>8</td><td>95</td></tr><tr><td>Pollen grains</td><td>RGB</td><td>unposed object</td><td>6601</td><td>12</td><td>99</td></tr><tr><td>Spiders</td><td>mono</td><td>genitalia</td><td>102</td><td>6</td><td>91</td></tr><tr><td>Human faces</td><td>mono</td><td>unposed face</td><td>400</td><td>41</td><td>99</td></tr></tbody></table>
<p>DAISY has been used in several research projects by O'Neill and others, and featured in popular science TV and magazine articles. The project has also been the subject of a recent article in <i><a href="/facts/Science_(journal)/uTyzjkwD">Science</a></i>.
</p><p>In 2011, the first DAISY installation capable of scaling to hundreds of taxa was installed at <a href="/facts/Natural_History_Museum%2c_London/SWrrNn0K">Natural History Museum</a> in London. This server offered both <a href="/facts/VNC/xTmWVgV4">VNC</a> and <a href="/facts/Web_service/qxrwqB3D">web service</a> based interfaces and was able to offload compute intensive pattern matching operations onto an <a href="/facts/NVIDIA/8QfCgqbu">NVIDIA</a> <a href="/facts/GPU/PTK1RQVp">GPU</a> programmed using <a href="/facts/CUDA/fvPpolot">CUDA</a>. This installation was capable of providing identification to species given  a 300+ taxon dataset in less than a second in a multiple user environment.
</p><p>More recently, under the aegis of <a href="/facts/Innovate_UK/zorRghE2">Innovate UK</a> funding, DAISY has been extensively modified to meet the needs of upstream activities within the oil and gas sector, in particular <a href="/facts/Biostratigraphy/AzdPX12G">biostratigraphy</a>. The resultant system, GeoDAISY represents a significant technological advance. It is capable of <a href="/facts/Deep_learning/JLuwD3ea">deep learning</a>, knowledge encapsulation, pattern based <a href="/facts/Data_mining/AH53q5ac">data mining</a> and <a href="/facts/Visual_search_engine/9IQmIa8e">(image based) content search</a> and can efficiently handle training sets consisting of millions of patterns on commodity hardware using a combination of smart data caching and <a href="/facts/OpenMP/69IpI80o">OpenMP</a>. Further details of GeoDAISY, and the rationale for developing it are available as white papers on the <a href="https://www.linkedin.com/company/tumbling-dice">Tumbling Dice LinkedIn page</a>.
</p>

Digital Automated Identification System open-in-new

Digital Automated Identification System