View Code? Open in Web Editor
NEW
data preprocess for cv, including data aug (general and cutout) for object detection, KITTI, COCO data to VOC
cv_toolbox's Introduction
DataAugForObjectDetection : 针对目标检测的增强脚本文件夹,增强方式包括
裁剪(会改变bbox)
平移(会改变bbox)
旋转(会改变bbox)
镜像(会改变bbox)
改变亮度
加噪声
cutout
KITTI_2_VOC : 将KITTI数据形式转换为VOC形式
VOC_2_COCO : 讲VOC形式数据转换为COCO格式
Attention : feature map Attention && channel attention (SE)
cv_toolbox's People
cv_toolbox's Issues
旋转后bbox到坐标没有跟着转呀,代码好像没做这样的处理
I get a data error when create data augment:
change DataAugmentForObjectDetection.py line351 :
flip_img = cv2.flip(flip_img, -1) to flip_img = cv2.flip(flip_img, 1)
Hello,I got a question to ask you.Since I am a novice, I have to ask you to tell me how to use it. Can you tell me?
DataAugmentForObejctDetection.py 446行将name字段过滤掉了
这与xml_helper.py中对于coords的要求是矛盾的