数据集格式：labelme格式(不包含mask文件，仅仅包含jpg图片和对应的json文件)
图片数量(jpg文件个数)：1015
标注数量(json文件个数)：1015
标注类别数：85
标注类别名称:"fabric01","fabric02","fabric03","fabric04","fabric05","fabric06","fabric07","fabric08","fabric09","fabric10","fabric11","fabric12","fabric13","fabric14","fabric15","fabric16","fabric17","fabric18","glass01","glass02","glass03","glass04","glass05","glass06","glass07","glass08","glass09","glass10","glass11","glass12","glass13","glass14","glass15","glass16","glass17","glass18","glass19","metal01","metal02","metal03","metal04","metal05","metal06","metal07","metal08","metal09","metal10","metal11","metal12","plastic01","plastic02","plastic03","plastic04","plastic05","plastic06","plastic07","plastic08","plastic09","plastic10","plastic11","plastic12","plastic13","plastic14","plastic15","po","wood01","wood02","wood03","wood04","wood05","wood06","wood07","wood08","wood09","wood10","wood11","wood12","wood13","wood14","wood15","wood16","wood17","wood18","wood19","wood20"
每个类别标注的框数和占有图片数：
fabric01框数=879,占有图片数=851
fabric02框数=697,占有图片数=685
fabric03框数=493,占有图片数=484
fabric04框数=348,占有图片数=343
fabric05框数=231,占有图片数=230
fabric06框数=156,占有图片数=155
fabric07框数=108,占有图片数=108
fabric08框数=73,占有图片数=73
fabric09框数=51,占有图片数=51
fabric10框数=48,占有图片数=47
fabric11框数=33,占有图片数=33
fabric12框数=21,占有图片数=21
fabric13框数=20,占有图片数=20
fabric14框数=9,占有图片数=9
fabric15框数=6,占有图片数=6
fabric16框数=3,占有图片数=3
fabric17框数=3,占有图片数=3
fabric18框数=1,占有图片数=1
glass01框数=562,占有图片数=559
glass02框数=422,占有图片数=420
glass03框数=264,占有图片数=264
glass04框数=179,占有图片数=179
glass05框数=118,占有图片数=118
glass06框数=71,占有图片数=71
glass07框数=44,占有图片数=44
glass08框数=35,占有图片数=35
glass09框数=22,占有图片数=22
glass10框数=18,占有图片数=18
glass11框数=13,占有图片数=13
glass12框数=10,占有图片数=10
glass13框数=8,占有图片数=8
glass14框数=6,占有图片数=6
glass15框数=3,占有图片数=3
glass16框数=3,占有图片数=3
glass17框数=3,占有图片数=3
glass18框数=2,占有图片数=2
glass19框数=1,占有图片数=1
metal01框数=264,占有图片数=261
metal02框数=137,占有图片数=134
metal03框数=63,占有图片数=63
metal04框数=34,占有图片数=34
metal05框数=23,占有图片数=23
metal06框数=13,占有图片数=13
metal07框数=5,占有图片数=5
metal08框数=5,占有图片数=5
metal09框数=4,占有图片数=4
metal10框数=4,占有图片数=4
metal11框数=3,占有图片数=3
metal12框数=1,占有图片数=1
plastic01框数=553,占有图片数=541
plastic02框数=260,占有图片数=257
plastic03框数=113,占有图片数=110
plastic04框数=62,占有图片数=59
plastic05框数=36,占有图片数=35
plastic06框数=21,占有图片数=21
plastic07框数=9,占有图片数=9
plastic08框数=6,占有图片数=6
plastic09框数=6,占有图片数=6
plastic10框数=2,占有图片数=2
plastic11框数=3,占有图片数=3
plastic12框数=2,占有图片数=2
plastic13框数=2,占有图片数=2
plastic14框数=2,占有图片数=2
plastic15框数=1,占有图片数=1
po框数=1,占有图片数=1
wood01框数=914,占有图片数=898
wood02框数=697,占有图片数=686
wood03框数=458,占有图片数=458
wood04框数=273,占有图片数=271
wood05框数=161,占有图片数=161
wood06框数=99,占有图片数=99
wood07框数=69,占有图片数=69
wood08框数=49,占有图片数=49
wood09框数=30,占有图片数=30
wood10框数=20,占有图片数=20
wood11框数=15,占有图片数=15
wood12框数=12,占有图片数=12
wood13框数=10,占有图片数=10
wood14框数=4,占有图片数=4
wood15框数=4,占有图片数=4
wood16框数=3,占有图片数=3
wood17框数=4,占有图片数=4
wood18框数=3,占有图片数=3
wood19框数=3,占有图片数=3
wood20框数=1,占有图片数=1
总框数：9428
使用标注工具：labelme=5.5.0
所在github仓库：firc-dataset
图片分辨率：512x512
标注规则：对类别进行画多边形框polygon
重要说明：可以将数据集用labelme打开编辑，json数据集需自己转成mask或者yolo格式或者coco格式作语义分割或者实例分割。针对室内场景构建易燃物分割数据集，按照燃烧特性把目标分为非常易燃（衣物、被褥）、易燃（木材、塑料）、不易燃（玻璃）三大类别。基于该数据集完成语义分割任务，能够有效衡量室内场景的可燃物危险程度。
特别声明：本数据集不对训练的模型或者权重文件精度作任何保证
图片预览：
标注例子：
原图（随机选16张图）：
标注绘制结果：
labelme编辑图实例：
单张图片：
