Note

I'm solving scale invariant. If you have a good paper, you can email me by [email protected]. Thanks!
Because I'm looking for an internship, I suspend update the project. If anyone is willing to recommend me to your company, thank you very much.
If you have any problem, please read README.md carefully

Improve yolo_v3 with latest paper

performance on VOC2007(Better performance than Tencent's reimplementation)

model	initial with yolov3.weights		initial with darknet.weights		release
	mAP	delta	mAP	delta
baseline	84.3	0.0	72.3	0.0	yes
data agumentation	85.8	+1.5	75.9	+3.6	yes
multi scale train	86.3	+0.5	78.3	+2.4	yes
focal loss	88.3	+2.0	79.6	+1.3	yes
des	xxx	+1.0	xxx	xxx	no
group normalization	xxx	xxx	xxx	xxx	yes
soft nms	xxx	-0.6	xxx	-0.6	yes
mix up	xxx	-0.3	80.7	+1.4	no
label smooth	88.6	+0.3	81.1	+0.4	yes
cosine learning rate	88.6	+0.0	82.1	+1.0	yes
GIOU	88.8	+0.2	82.3	+0.2	yes
remove anchor(running)	xxx	xxx	xxx	xxx	no
multi scale test	90.7	+1.9	84.1	+1.8	yes

evaluated at 544x544 on Pascal VOC 2007 test set

to do

• Deformable convolutional networks
  
• Scale-Aware Trident Networks for Object Detection
• Understanding the Effective Receptive Field in Deep Convolutional Neural Networks
  

Usage

1. clone YOLO_v3 repository

   git clone https://github.com/Stinky-Tofu/Stronger-yolo.git

2. prepare data
   (1) download datasets
   Create a new folder named data in the directory where the YOLO_V3 folder is located, and then create a new folder named VOC in the data/.
   Download VOC 2012_trainval 、VOC 2007_trainval 、VOC 2007_test, and put datasets into data/VOC, name as 2012_trainval、2007_trainval、2007_test separately.
   The file structure is as follows:
   |--YOLO_V3
   |--data
   |--|--VOC
   |--|--|--2012_trainval
   |--|--|--2007_trainval
   |--|--|--2007_test
   (2) convert data format
   You should set DATASET_PATH in config.py to the path of the VOC dataset, for example: DATASET_PATH = '/home/xzh/doc/code/python_code/data/VOC',and then
   

   python voc_annotation.py

3. prepare initial weights
   Download YOLOv3-608.weights firstly, put the yolov3.weights into yolov3_to_tf/, and then

   cd yolov3_to_tf
   python3 convert_weights.py --weights_file=yolov3.weights --dara_format=NHWC --ckpt_file=./saved_model/yolov3_608_coco_pretrained.ckpt
   cd ..
   python rename.py

4. Train
   

   python train.py

5. Test
   Download weight file yolo_test.ckpt
   If you want to get a higher mAP, you can set the score threshold to 0.01、use multi scale test、flip test.
   If you want to use it in actual projects, or if you want speed, you can set the score threshold to 0.2.
   

   python test.py --gpu=0 --map_calc=True --weights_file=model_path.ckpt
   cd mAP
   python main.py -na -np

Train for custom dataset

1. Generate your own annotation file train_annotation.txt and test_annotation.txt, one row for one image.
   Row format: image_path bbox0 bbox1 ...
   Bbox format: xmin,ymin,xmax,ymax,class_id(no space), for example:
   

   /home/xzh/doc/code/python_code/data/VOC/2007_test/JPEGImages/000001.jpg 48,240,195,371,11 8,12,352,498,14

2. Put the train_annotation.txt and test_annotation.txt into YOLO_V3/data/.
   
3. Configure config.py for your dataset.
   
4. Start training.
   

   python train.py

Reference:

paper:

• YOLOv3: An Incremental Improvement
  
• Foca Loss for Dense Object Detection
  
• Group Normalization
  
• Single-Shot Object Detection with Enriched Semantics
  
• An Analysis of Scale Invariance in Object Detection - SNIP
  
• Deformable convolutional networks
  
• Scale-Aware Trident Networks for Object Detection
  
• Understanding the Effective Receptive Field in Deep Convolutional Neural Networks
  
• Bag of Freebies for Training Object Detection Neural Networks
  
• Generalized Intersection over Union: A Metric and A Loss for Bounding Box Regression
  

mAP calculate: mean Average Precision

Requirements

software

• Python2.7.12
  
• Numpy1.14.5
  
• Tensorflow.1.8.0
  
• Opencv3.4.1
  

headware

• 12G 1080Ti