注意这个是识别牙齿的不包含任何牙齿疾病在里面，注意查看图片

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)：3971

标注数量(json文件个数)：3971

标注类别数：32

标注类别名称:["LL1","LL2","LL3","LL4","LL5","LL6","LL7","LL8","LR1","LR2","LR3","LR4","LR5","LR6","LR7","LR8","UL1","UL2","UL3","UL4","UL5","UL6","UL7","UL8","UR1","UR2","UR3","UR4","UR5","UR6","UR7","UR8"]

每个类别标注的框数：

LL1 count = 3926

LL2 count = 3943

LL3 count = 3953

LL4 count = 3876

LL5 count = 3781

LL6 count = 3480

LL7 count = 3702

LL8 count = 2602

LR1 count = 3930

LR2 count = 3924

LR3 count = 3952

LR4 count = 3890

LR5 count = 3800

LR6 count = 3493

LR7 count = 3712

LR8 count = 2581

UL1 count = 3879

UL2 count = 3828

UL3 count = 3882

UL4 count = 3714

UL5 count = 3663

UL6 count = 3678

UL7 count = 3783

UL8 count = 2488

UR1 count = 3877

UR2 count = 3837

UR3 count = 3878

UR4 count = 3713

UR5 count = 3691

UR6 count = 3703

UR7 count = 3776

UR8 count = 2420

总框数：116355

使用标注工具：labelme=5.5.0

所在github仓库：firc-dataset

图片分辨率：640x640

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

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：