Giter Club home page Giter Club logo

detector-in-detector's Introduction


Detector-in-Detector

“Detector-in-Detector: Multi-Level Analysis for Human-Parts” 的非官方实现:针对人体部位的多层次分析。

In English

简要引述

该项目基于TensorFlow目标检测框架Luminoth,并重载了Faster-RCNN接口,以实现简化形式的 Detector-in-Detector。

论文链接: https://arxiv.org/abs/1902.07017

安装

train.py
models/fasterrcnn/__init__.py
models/fasterrcnn/detector_in_detector.py
models/fasterrcnn/part_detector.py
models/fasterrcnn/part_detector_rcnn.py
models/fasterrcnn/rcnn_overload.py
models/fasterrcnn/rcnn_proposal_overload.py
models/fasterrcnn/part_detector_roi_pool.py
predict.py
utils/detector_in_detector_predicting.py

训练过程

  • 1 根据下面的链接下载论文中的数据集
    https://github.com/xiaojie1017/Human-Parts
  • 2 根据 dataset_script/preprocess_dataset.py 来生成适合 luminoth
    使用的csv标注文件
  • 3 根据 luminoth dataset api 生成 tfrecord 格式数据文件
  • 4 使用 luminoth train -c dataset_script/start.yml 来训练
  • 5 使用 tensorboard 来观察训练过程.
  • 6 使用 luminoth predict -c dataset_script/start.yml 来预测

细节

如何调试源码

在训练过程中:
编辑 utils/hooks/image_vis_hook.py 从挂钩函数 after_run 中获得结果

在预测过程中:
编辑 utils/detector_in_detector_predicting.py

与 luminoth 的一些区别

这只是luminoth FasterRCNN 的编辑更改版本。与原版不同之处在于将 main_label_index 设置为区分主体(main_part 的)bbox和 身体组成成分的 bbox。

与论文实现叙述的区别

论文中使用人物的基准特征作为 padding 来进行早期训练过程,这里则被随机采样和概率阈值过滤替换。您可以尝试增加填充数量(即7)来逼近它。或者您也可以手动切换。

一些可能需要更改的实现缺陷

  • 1 对于过滤脚本(filter script)的使用,数据增强配置仅限于翻转和随机噪声添加动作,有些数据增强操作可能会导致主要和部分的重叠为零,这对训练是有害的。这与身体(body)检测器和部件(part)检测器之间的构造有关,如果您可以积累样本并筛选有效样本的方式进行训练,您可以解决这个问题。

  • 2 luminoth faster-rcnn 只支持 batch-size 1, 限制了训练效率.

  • 3 没有重载 luminoth eval 和 网络服务部署接口 (较为简单)

Contact

svjack - [email protected] - [email protected]

Project Link:https://github.com/svjack/Detector-in-Detector

detector-in-detector's People

Contributors

svjack avatar

Stargazers

 avatar  avatar  avatar

Watchers

 avatar

Forkers

hoangkhoile

detector-in-detector's Issues

Pretrained model

Hello,

Thank you for your excellent work. Can you provide the pretrained model ?

Best regards

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.