yolov1-resnet

YOLO: Real-Time Object Detection

Train on voc2012+2007

Model	Backbone	mAP@voc2007test	FPS
ResNet YOLOv1	ResNet50	0.49	__
YOLOv1	darknet19	0.634	45

Dataset:

1. Download voc2012train dataset
2. Download voc2007train dataset
3. Download voc2007test dataset
4. Put all images in JPEGImages folder in voc2012train and voc2007train to Images folder as following:

   ├── Dataset 
    ├── Images
        ├── 0001.jpg
        ├── 0002.jpg
    ├── Labels
        ├── 0001.txt
        ├── 0002.txt
    ├── train.txt
    ├── test.txt
   

Each label consists of class and bounding box information. e.g 0001.txt :

1 255 247 425 468
0 470 105 680 468
1 152 356 658 754

How to convert .xml files to .txt format?

• Download this repo and modify config.py to convert VOC format to YOLO format labels

Train:

• python main.py

Evaluation:

• python eval.py
• In evaluation.py, im_show=False change to True to see the results.

Detection:

• To show the result image - python detect.py --image assets/person.jpg
• Save result image - python detect.py --image assets/person.jpg --save_img

Weights:

• Run the download.sh file in weights folder or download from this link

Result:

CLASS                     AP
aeroplane                 0.54
bicycle                   0.60
bird                      0.48
boat                      0.28
bottle                    0.16
bus                       0.63
car                       0.59
cat                       0.71
chair                     0.24
cow                       0.48
diningtable               0.39
dog                       0.64
horse                     0.58
motorbike                 0.59
person                    0.51
pottedplant               0.17
sheep                     0.47
sofa                      0.50
train                     0.69
tvmonitor                 0.46
mAP: 0.49

any questions are welcome…

Reference:

• https://github.com/abeardear/pytorch-YOLO-v1