TinySSD-windows-ssd-caffe

Windows Setup

• install caffe-ssd-windows copy the file create_annoset.py to caffe-ssd-windows-master/scripts
• install anaconda
• downlaod voc datasets

get TinySSD-windows-ssd-caffe

• download this project

get the person dataset from the standard voc datasets

• run extra_voc_person/extra_person_2007.py and extra_voc_person/extra_person_2012.py to extra the person data list
• run extra_voc_person to extra the name of person from xml and delete the other name object, then copy the xml to person_Annotations dirctionary
• run voc/create_list.sh and get_image_size.bat to create the text list of person image and xml file path
• run create_data.sh to create the lmdb datasets

train

• run TinySSD-windows-ssd-caffe/train.bat