Git Product home page Git Product logo

centernet-pytorch's Introduction

CenterNet:Objects as Points目标检测模型在Keras当中的实现


目录

  1. 性能情况 Performance
  2. 所需环境 Environment
  3. 注意事项 Attention
  4. 文件下载 Download
  5. 预测步骤 How2predict
  6. 训练步骤 How2train
  7. 参考资料 Reference

性能情况

训练数据集 权值文件名称 测试数据集 输入图片大小 mAP 0.5:0.95 mAP 0.5
VOC07+12 centernet_resnet50_voc.pth VOC-Test07 512x512 - 77.1
COCO-Train2017 centernet_hourglass_coco.pth COCO-Val2017 512x512 38.4 56.8

所需环境

torch==1.2.0

注意事项

代码中的centernet_resnet50_voc.pth是使用voc数据集训练的。
代码中的centernet_hourglass_coco.pth是使用voc数据集训练的。
注意不要使用中文标签,文件夹中不要有空格!
在训练前需要务必在model_data下新建一个txt文档,文档中输入需要分的类,在train.py中将classes_path指向该文件

文件下载

训练所需的centernet_resnet50_voc.pth、centernet_hourglass_coco.pth可在百度网盘中下载。
链接: https://pan.baidu.com/s/1QBBgRb_TH8kJdSCQGgcXmQ 提取码: phnc
centernet_resnet50_voc.pth是voc数据集的权重。
centernet_hourglass_coco.pth是coco数据集的权重。

预测步骤

1、使用预训练权重

a、下载完库后解压,在百度网盘下载centernet_resnet50_voc.pth或者centernet_hourglass_coco.pth,放入model_data,运行predict.py,输入

img/street.jpg

可完成预测。
b、利用video.py可进行摄像头检测。

2、使用自己训练的权重

a、按照训练步骤训练。
b、在yolo.py文件里面,在如下部分修改model_path和classes_path使其对应训练好的文件;model_path对应logs文件夹下面的权值文件,classes_path是model_path对应分的类

_defaults = {
    "model_path"        : 'model_data/centernet_resnet50_voc.pth',
    "classes_path"      : 'model_data/voc_classes.txt',
    # "model_path"        : 'model_data/centernet_hourglass_coco.h5',
    # "classes_path"      : 'model_data/coco_classes.txt',
    "backbone"          : "resnet50",
    "image_size"        : [512,512,3],
    "confidence"        : 0.3,
    # backbone为resnet50时建议设置为True
    # backbone为hourglass时建议设置为False
    # 也可以根据检测效果自行选择
    "nms"               : True,
    "nms_threhold"      : 0.3,
    "cuda"              : True
}

c、运行predict.py,输入

img/street.jpg

可完成预测。
d、利用video.py可进行摄像头检测。

训练步骤

1、本文使用VOC格式进行训练。
2、训练前将标签文件放在VOCdevkit文件夹下的VOC2007文件夹下的Annotation中。
3、训练前将图片文件放在VOCdevkit文件夹下的VOC2007文件夹下的JPEGImages中。
4、在训练前利用voc2centernet.py文件生成对应的txt。
5、再运行根目录下的voc_annotation.py,运行前需要将classes改成你自己的classes。注意不要使用中文标签,文件夹中不要有空格!

classes = ["aeroplane", "bicycle", "bird", "boat", "bottle", "bus", "car", "cat", "chair", "cow", "diningtable", "dog", "horse", "motorbike", "person", "pottedplant", "sheep", "sofa", "train", "tvmonitor"]

6、此时会生成对应的2007_train.txt,每一行对应其图片位置及其真实框的位置
7、在训练前需要务必在model_data下新建一个txt文档,文档中输入需要分的类,在train.py中将classes_path指向该文件,示例如下:

classes_path = 'model_data/new_classes.txt'    

model_data/new_classes.txt文件内容为:

cat
dog
...

8、运行train.py即可开始训练。

mAP目标检测精度计算更新

更新了get_gt_txt.py、get_dr_txt.py和get_map.py文件。
get_map文件克隆自https://github.com/Cartucho/mAP
具体mAP计算过程可参考:https://www.bilibili.com/video/BV1zE411u7Vw

Reference

https://github.com/xuannianz/keras-CenterNet
https://github.com/see--/keras-centernet
https://github.com/xingyizhou/CenterNet

centernet-pytorch's People

Contributors

bubbliiiing avatar

Watchers

 avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.