Focal Loss for Dense Rotation Object Detection

Abstract

This repo is based on Focal Loss for Dense Object Detection, and it is completed by YangXue.

We also recommend a tensorflow-based rotation detection benchmark, which is led by YangXue.

Performance

DOTA1.0

Model	Backbone	Training data	Val data	mAP	Model Link	Anchor	Reg. Loss	Angle Range	lr schd	Data Augmentation	GPU	Image/GPU	Configs
RetinaNet	ResNet50_v1d 600->800	DOTA1.0 trainval	DOTA1.0 test	63.18	model	H	smooth L1	90	1x	No	1X GeForce RTX 2080 Ti	1	cfgs_res50_dota_v4.py
RetinaNet	ResNet50_v1d 600->800	DOTA1.0 trainval	DOTA1.0 test	64.10	-	H	smooth L1	180	1x	No	1X GeForce RTX 2080 Ti	1	cfgs_res50_dota_v15.py

RetinaNet	ResNet50_v1d 600->800	DOTA1.0 trainval	DOTA1.0 test	62.76	model	R	smooth L1	90	1x	No	1X GeForce RTX 2080 Ti	1	cfgs_res50_dota_v1.py
RetinaNet	ResNet50_v1d 600->800	DOTA1.0 trainval	DOTA1.0 test	68.65	-	R	iou-smooth L1	90	1x	No	1X GeForce RTX 2080 Ti	1	cfgs_res50_dota_v5.py

Notice:
Due to the improvement of the code, the performance of this repo is gradually improving, so the experimental results in configuration files are for reference only.
This repo has basically stopped updating, please refer to new repo for the latest progress.

Visualization

My Development Environment

docker images: docker pull yangxue2docker/yx-tf-det:tensorflow1.13.1-cuda10-gpu-py3
1、python3.5 (anaconda recommend)
2、cuda 10.0
3、opencv(cv2)
4、tfplot 0.2.0 (optional)
5、tensorflow 1.13

IoU-smooth L1 Loss

SCRDet: Towards More Robust Detection for Small, Cluttered and Rotated Objects (ICCV2019)

Download Model

Pretrain weights

1、Please download resnet50_v1, resnet101_v1 pre-trained models on Imagenet, put it to data/pretrained_weights.
2、(Recommend in this repo) Or you can choose to use a better backbone, refer to gluon2TF.

Baidu Drive, password: 5ht9.
Google Drive

Compile

cd $PATH_ROOT/libs/box_utils/cython_utils
python setup.py build_ext --inplace (or make)

cd $PATH_ROOT/libs/box_utils/
python setup.py build_ext --inplace

Train

1、If you want to train your own data, please note:

(1) Modify parameters (such as CLASS_NUM, DATASET_NAME, VERSION, etc.) in $PATH_ROOT/libs/configs/cfgs.py
(2) Add category information in $PATH_ROOT/libs/label_name_dict/label_dict.py     
(3) Add data_name to $PATH_ROOT/data/io/read_tfrecord.py

2、Make tfrecord
For DOTA dataset:

cd $PATH_ROOT\data\io\DOTA
python data_crop.py

cd $PATH_ROOT/data/io/  
python convert_data_to_tfrecord.py --VOC_dir='/PATH/TO/DOTA/' 
                                   --xml_dir='labeltxt'
                                   --image_dir='images'
                                   --save_name='train' 
                                   --img_format='.png' 
                                   --dataset='DOTA'

3、Multi-gpu train

cd $PATH_ROOT/tools
python multi_gpu_train.py

Test

cd $PATH_ROOT/tools
python test_dota.py --test_dir='/PATH/TO/IMAGES/'  
                    --gpus=0,1,2,3,4,5,6,7

Notice: In order to set the breakpoint conveniently, the read and write mode of the file is' a+'. If the model of the same #VERSION needs to be tested again, the original test results need to be deleted.

Tensorboard

cd $PATH_ROOT/output/summary
tensorboard --logdir=.

Reference

1、https://github.com/endernewton/tf-faster-rcnn
2、https://github.com/zengarden/light_head_rcnn
3、https://github.com/tensorflow/models/tree/master/research/object_detection
4、https://github.com/fizyr/keras-retinanet

DetectionTeamUCAS / RetinaNet_Tensorflow_Rotation