i_naturalist2017

इस डेटासेट में 579,184 प्रशिक्षण छवियों और 95,986 सत्यापन छवियों में कुल 5,089 श्रेणियां हैं। प्रशिक्षण सेट के लिए, प्रति श्रेणी छवियों का वितरण iNaturalist समुदाय द्वारा उस श्रेणी की अवलोकन आवृत्ति का अनुसरण करता है।

हालांकि मूल डेटासेट में बाउंडिंग बॉक्स वाली कुछ छवियां हैं, वर्तमान में, केवल छवि-स्तरीय एनोटेशन (एकल लेबल/छवि) प्रदान किए जाते हैं। इसके अलावा, आयोजकों ने परीक्षण लेबल प्रकाशित नहीं किए हैं, इसलिए हम केवल परीक्षण चित्र प्रदान करते हैं (लेबल = -1)।

विभाजित करना उदाहरण
'test' 182,707
'train' 579,184
'validation' 95,986
  • फ़ीचर संरचना :
FeaturesDict({
    'id': Text(shape=(), dtype=string),
    'image': Image(shape=(None, None, 3), dtype=uint8),
    'label': ClassLabel(shape=(), dtype=int64, num_classes=5089),
    'supercategory': ClassLabel(shape=(), dtype=int64, num_classes=13),
})
  • फ़ीचर दस्तावेज़ीकरण :
विशेषता कक्षा आकार डीटाइप विवरण
विशेषताएं डिक्ट
पहचान मूलपाठ डोरी
छवि छवि (कोई नहीं, कोई नहीं, 3) uint8
लेबल क्लासलेबल int64
superश्रेणी क्लासलेबल int64

VISUALIZATION

  • उद्धरण :
@InProceedings{Horn_2018_CVPR,
author = {
Van Horn, Grant and Mac Aodha, Oisin and Song, Yang and Cui, Yin and Sun, Chen
and Shepard, Alex and Adam, Hartwig and Perona, Pietro and Belongie, Serge},
title = {The INaturalist Species Classification and Detection Dataset},
booktitle = {
The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2018}
}