Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
A deep learning based algorithm to detect rotated object, for example, objects in remote sensing images
| Date | Stars |
|---|---|
| 2026-07-31 | 426 |
| 2026-08-04 | 426 |
| 2026-08-06 | 426 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
# DRBox By Lei Liu (mail: [email protected]) ### Introduction DRBox is used for detection tasks where the objects are orientated arbitrarily. This code show examples that DRBox is used to detect vehicles, ships and airplanes in remote sensing images. I'm also looking forward for its use in other problems.  The codes are modified from the original Caffe and [SSD](https://github.com/weiliu89/caffe/tree/ssd). ### Citing DRBox The article for this method can be downloaded here: [arXiv:1711.09405](https://arxiv.org/abs/1711.09405). Please cite this work in your publications if it helps your research. ### Contents 1. [Installation](#installation) 2. [Preparation](#preparation) 3. [Train](#train) 4. [Deployment](#deployment) 5. [View Results](#view-results) 6. [Build Your Own Dataset](#build-your-own-dataset) ### Installation 1. DRBox is written in Caffe with some newly defined layers. So you should prepare nessasary environment for Caffe installation. DThen you can get the code: ```Shell git clone https://github.come/liulei01/drbox.git ``` 2. Matlab is also neccessary so that the results can be viewed. 3. If you only want to apply our trained models directly to your applications, then you can ignore the following instruction and jump to [Deployment](#deployment). 4. Build the code. Please follow [Caffe instruction](http://caffe.berkeleyvision.org/installation.html) to install all necessary packages and build it. ```Shell # Modify Makefile.config according to your Caffe installation. cp Makefile.config.example Makefile.config make -j8 # Make sure to include $CAFFE_ROOT/python to your PYTHONPATH. make py ``` ### Preparation 1. Download [fully convolutional reduced (atrous) VGGNet](https://gist.github.com/weiliu89/2ed6e13bfd5b57cf81d6). By default, we assume the model is stored in `$CAFFE_ROOT/models/VGGNet/` 2. Download the training data for DRBox at https://pan.baidu.com/s/1sliHG09 , extract them and place the files at the corresponding position of data/. (Tips: There are two folders named larger and smaller respectively, you can download files in either of them. This website would ask you to install a software to download files that larger than 200M, so if you want to download file from your browser directly, you should download the files in folder "smaller".) 3. Run create_data.sh in each subfolders in data/ to create LMDB for training. For example, when you want to train a airplane detection network, then you can ```Shell cd $CAFFEROOT ./data/Airplane/create_data.sh ``` ### Train DRBox is now designed as a single task network. So you should train it for each type of objects separately. The python codes are in examples/rbox/. If you want to train a airplane detection network, then you can start training by: ```Shel cd $CAFFEROOT python examples/rbox/rbox_pascal_airplane.py ``` Training for vehicle is similar with airplane. ```Shell cd $CAFFEROOT python examples/rbox/rbox_pascal_car.py ``` Before training for ship, you should replace src/caffe/util/rbox_util.cpp with src/caffe/util.rbox_util.cpp.ship and rebuilding the codes. The reason is that we ignore the head and tail of a ship to make the problem easier. ```Shell cd $CAFFEROOT mv src/caffe/util/rbox_util.cpp src/caffe/util/rbox_util.cpp.old mv src/caffe/util/rbox_util.cpp.ship src/caffe/util/rbox_util.cpp make -j8 python examples/rbox/rbox_pascal_ship_opt.py ``` The trained models are stored in models/RBOX/. ### Deployment The codes for deployment are in examples/rbox/deploy. 1. If you only want to apply a pre-trained models directly to your applications in any Caffe environment, then you can copy this folder to your own Caffe folder and run the following commands. ```Shell mv librbox.cpp.code librbox.cpp g++ -o librbox.so -shared -fPIC librbox.cpp cp deploy.py.general_example deploy.py # you should modify the following file accordingly. python deploy.py ``` 2. Otherwise, make sure that
Excerpt of 6,020 characters
Read on GitHubWould you bet a product on this? Bounded 0–100 and slow moving.
matched fp:011bf34cf006181b, llm:Description, README: 'A deep learning based algorithm to detect rotated object... examples that DRBox is used to detect vehicles, ships and airplanes in remote sensing images.' Topics: caffe, detection, rotation-invariant. Uses modified Caffe and SSD for rotated object detection in remote sensing.
matched fp:011bf34cf006181b, llm:Description, README: 'A deep learning based algorithm to detect rotated object... examples that DRBox is used to detect vehicles, ships and airplanes in remote sensing images.' Topics: caffe, detection, rotation-invariant. Uses modified Caffe and SSD for rotated object detection in remote sensing.
matched fp:011bf34cf006181b, llm:Description, README: 'A deep learning based algorithm to detect rotated object... examples that DRBox is used to detect vehicles, ships and airplanes in remote sensing images.' Topics: caffe, detection, rotation-invariant. Uses modified Caffe and SSD for rotated object detection in remote sensing.