Samples
Overview
Openlogo是一个用于检测logo的数据集,数据集共包含有352种logo(如下),27924张图片,其中含annotation的图片有27083张
'guinness', 'dhl', 'stellaartois', 'fosters', 'becks', 'nvidia', 'shell', 'corona', 'ford', 'google', 'aldi', 'pepsi', 'apple', 'fedex', 'starbucks', 'singha', 'heineken', 'carlsberg', 'paulaner', 'ferrari', 'ups',
'cocacola', 'erdinger', 'tsingtao', 'texaco', 'esso', 'chimay', 'bmw', 'adidas', 'rittersport', 'hp', 'milka', 'adidas_text', '3m', 'heineken_text', 'abus', 'accenture', 'adidas1', 'airhawk', 'honda_text', 'aldi_text',
'target_text', 'walmart_text', 'subway', 'amazon', 'alfaromeo', 'allett', 'allianz', 'allianz_text', 't-mobile', 'aluratek', 'aluratek_text', 'playstation', 'amcrest', 'amcrest_text', 'americanexpress', 'americanexpress_text',
'mastercard', 'visa', 'android', 'anz', 'anz_text', 'apc', 'apecase', 'aquapac_text', 'aral', 'audi', 'mcdonalds', 'mercedesbenz', 'lotto', 'volkswagen', 'honda', 'ec', 'armani', 'armitron', 'aspirin', 'bayer', 'asus', 'athalon',
'audi_text', 'michelin', 'bosch_text', 'mercedesbenz_text', 'lexus', 'axa', 'bacardi', 'bankofamerica', 'bankofamerica_text', 'kodak', 'samsung', 'barbie', 'barclays', 'basf', 'batman', 'bbc', 'bbva', 'tnt', 'umbro',
'standard_liege', 'gucci', 'dexia', 'puma', 'puma_text', 'nike', 'base', 'citroen_text', 'total', 'bfgoodrich', 'kia', 'citroen', 'quick', 'carglass', 'airness', 'bik', 'reebok1', 'bridgestone_text', 'us_president',
'bridgestone', 'bellataylor', 'bellodigital', 'bellodigital_text', 'bem', 'twitter', 'youtube', 'benrus', 'bershka', 'bionade', 'blackmores', 'blizzardentertainment', 'volvo', 'mini', 'toyota', 'facebook', 'intel',
'teslamotors', 'boeing', 'boeing_text', 'unicef', 'bosch', 'porsche', 'porsche_text', 'opel', 'siemens', 'panasonic', 'bottegaveneta', 'budweiser', 'budweiser_text', 'corona_text', 'bulgari', 'burgerking',
'burgerking_text', 'pizzahut', 'costa', 'mcdonalds_text', 'calvinklein', 'canon', 'sony', 'carters', 'cartier', 'caterpillar', 'chanel', 'chanel_text', 'prada', 'chevrolet', 'chevrolet_text', 'subaru', 'chevron', 'chickfila',
'chiquita', 'cisco', 'rbc', 'citi', 'coach', 'coke', 'sprite', 'gap', 'hyundai', 'colgate', 'comedycentral', 'converse', 'costco', 'homedepot_text', 'cpa_australia', 'cvs', 'cvspharmacy', 'danone', 'disney', 'drpepper',
'dunkindonuts', 'ebay', 'espn', 'esso_text', 'mobil', 'tigerwash', 'spar', 'spar_text', 'shell_text', 'renault', 'evernote', 'nissan', 'firefox', 'nbc', 'redbull', 'vodafone', 'yahoo', 'fly_emirates', 'fritolay', 'fritos',
'cheetos', 'doritos', 'lays', 'ruffles', 'sunchips', 'tostitos', 'pepsi_text', 'lg', 'generalelectric', 'gildan', 'gillette', 'venus', 'walmart', 'goodyear', 'hanes', 'head', 'head_text', 'heraldsun', 'hermes', 'hersheys',
'kitkat', 'reeses', 'hh', 'hisense', 'hm', 'jcrew', 'homedepot', 'hsbc', 'hsbc_text', 'rolex', 'huawei_text', 'huawei', 'sap', 'hyundai_text', 'ikea', 'ibm', 'at_and_t', 'windows', 'internetexplorer', 'jackinthebox',
'jacobscreek', 'jagermeister', 'johnnywalker', 'jurlique', 'kelloggs', 'lego', 'kfc', 'suzuki', 'kraft', 'jello', 'maxwellhouse', 'miraclewhip', 'philadelphia', 'planters', 'velveeta', 'lacoste', 'lacoste_text', 'lamborghini',
'levis', 'lexus_text', 'londonunderground', 'loreal', 'lv', 'luxottica', 'marlboro', 'marlboro_text', 'marlboro_fig', 'maserati', 'maxxis', 'mccafe', 'philips', 'medibank', 'microsoft', 'millerhighlife', 'mitsubishi',
'mk', 'motorola', 'mtv', 'nasa', 'nb', 'nescafe', 'netflix', 'nike_text', 'nintendo', 'nissan_text', 'infiniti', 'infiniti_text', 'nivea', 'northface', 'obey', 'olympics', 'optus', 'optus_yes', 'oracle', 'pampers',
'pepsi_text1', 'pizzahut_hut', 'poloralphlauren', 'recycling', 'redbull_text', 'firelli', 'santander_text', 'reebok_text', 'rolex_text', 'reebok', 'tommyhilfiger', 'republican', 'zara', 'santander', 'schwinn', 'sega',
'shell_text1', 'select', 'skechers', 'vaio', 'soundcloud', 'soundrop', 'spiderman', 'superman', 'supreme', 'tacobell', 'target', 'toyota_text', 'thomsonreuters', 'timberland', 'tissot', 'verizon', 'scion_text',
'underarmour', 'uniqlo', 'uniqlo1', 'unitednations', 'verizon_text', 'volkswagen_text', 'warnerbros', 'wellsfargo', 'wellsfargo_text', 'wii', 'williamhill', 'wordpress', 'xbox', 'yamaha', 'yonex_text', 'yonex'
Data Explore
Annotations -- 27083个xml格式的box2d标注文件,其格式如下
<annotation>
<folder>training</folder>
<filename>3m4.jpg</filename>
<source>
<database>synthetic initialization</database>
<annotation>PASCAL VOC2007</annotation>
<image>synthetic</image>
<flickrid>none</flickrid>
</source>
<owner>
<flickrid>none</flickrid>
<name>none</name>
</owner>
<size>
<width>1000</width>
<height>662</height>
<depth>3</depth>
</size>
<segmented>0</segmented>
<object>
<name>3m</name>
<pose>Unspecified</pose>
<truncated>0</truncated>
<difficult>0</difficult>
<bndbox>
<xmin>257</xmin>
<ymin>136</ymin>
<xmax>343</xmax>
<ymax>195</ymax>
</bndbox>
</object>
</annotation>
ImageSets -- 官方推荐的训练集、验证集、测试集划分
JPEGImages -- 27924张logo原图
数据预处理
输入图片对应的的xml文件路径,即可获得图片的box2D标签,其格式为[xmin, ymin, xmax, ymax, category]
def acquire_label_xml(img_path):
DOMTree = xml.dom.minidom.parse(img_path)
collection = DOMTree.documentElement
boundingbox = collection.getElementsByTagName("object")
img_lable = []
for i in boundingbox:
tmp = []
category = i.getElementsByTagName("name")[0].childNodes[0].data
tmp.append(float(
[j.childNodes[0].data for j in i.getElementsByTagName("bndbox")[0].getElementsByTagName("xmin")][0]))
tmp.append(float(
[j.childNodes[0].data for j in i.getElementsByTagName("bndbox")[0].getElementsByTagName("ymin")][0]))
tmp.append(float(
[j.childNodes[0].data for j in i.getElementsByTagName("bndbox")[0].getElementsByTagName("xmax")][0]))
tmp.append(float(
[j.childNodes[0].data for j in i.getElementsByTagName("bndbox")[0].getElementsByTagName("ymax")][0]))
tmp.append(category)
img_lable.append(tmp)
return img_lable
Citation
Open Logo Detection Challenge.
Hang Su, Xiatian Zhu and Shaogang Gong.
In Proc. British Machine Vision Conference (BMVC), Newcastle, UK, September 2018.
BibTex arxiv
关注公众号,回复 openlogo 即可获得数据集