SSD_scene_text_detection
SSD_scene_text_detection copied to clipboard
Detect text in natural images with SSD, Single Shot Detection
想请问下你这边训练时正负样本比例,以及产生正样本的阈值及loss的权重
I have got the real sizes and aspect ratios of the objects, how can I use them in the mxnet? Assuming: num_classes = 6 width: 29.55, 41.71, 72.41, 82.82, 96.68,...
I am run it to train, but error happened in this step: solver = caffe.SGDSolver(caffe_root + 'models/VGGNet/ICDAR2017/SSD_300x300/solver.prototxt') the trace stack is : F0708 18:54:26.167244 20822 # **math_functions.cpp:250] Check failed: a
/home/yang/caffe/build/tools/convert_annoset --anno_type=detection --label_type=xml --label_map_file=/home/yang/data/scenetext/labelmap_voc.prototxt --check_label=True --min_dim=0 --max_dim=0 --resize_height=0 --resize_width=0 --backend=lmdb --shuffle=False --check_size=False --encode_type=jpg --encoded=True --gray=False /home/yang/data/VOCdevkit/ /home/yang/data/scenetext/test.txt /home/yang/data/VOCdevkit/scenetext/lmdb/scenetext_test_lmdb Traceback (most recent call last): File "/home/yang/caffe/scripts/create_annoset.py", line 159, in process =...
When I run the `create_lmdbdata.sh`, I get two errors related to the `.xml` files in train and test folders. it says: ``` annotation file: /home/hesam/Desktop/img/val/annot/0269.xml does not exist annotation file:...
请问一下作者:你这个文件当中的trainval_dir以及test_dir,这两个路径当中的文件是ICDAR2011这个数据集当中的文件吗?如果是的话,也没有test以及trainval这个文件夹啊,麻烦说明一下如何修改你的代码当中的路径,谢谢了
error run create_data.sh , only generate the Soft connection,the size is 57 bytes , but did not genereate the actual lmdb . this is the result of running the create_data.sh...
trainval_dir = "/home/chenxp/data/VOCdevkit/scenetext/trainval" ??? What is VOCdevkit / scenetext?