All Projects → yuxitong → Tensorflowandroiddemo

yuxitong / Tensorflowandroiddemo

TensorFlow android demo 车道线 车辆 人脸 动作 骨架 识别 检测 抽烟 打电话 闭眼 睁眼

Programming Languages

java
68154 projects - #9 most used programming language

Projects that are alternatives of or similar to Tensorflowandroiddemo

Imodels
Interpretable ML package 🔍 for concise, transparent, and accurate predictive modeling (sklearn-compatible).
Stars: ✭ 194 (-67.06%)
Mutual labels:  artificial-intelligence, ai, demo
Text summurization abstractive methods
Multiple implementations for abstractive text summurization , using google colab
Stars: ✭ 359 (-39.05%)
Mutual labels:  artificial-intelligence, ai
Deeppavlov
An open source library for deep learning end-to-end dialog systems and chatbots.
Stars: ✭ 5,525 (+838.03%)
Mutual labels:  artificial-intelligence, ai
The Math Of Intelligence
List of resources & possible pathway for the Math of Machine Learning and AI.
Stars: ✭ 370 (-37.18%)
Mutual labels:  artificial-intelligence, ai
Csinva.github.io
Slides, paper notes, class notes, blog posts, and research on ML 📉, statistics 📊, and AI 🤖.
Stars: ✭ 342 (-41.94%)
Mutual labels:  artificial-intelligence, ai
Intelligo
🤖 Chatbot Framework for Node.js.
Stars: ✭ 347 (-41.09%)
Mutual labels:  artificial-intelligence, ai
Mycroft Core
Mycroft Core, the Mycroft Artificial Intelligence platform.
Stars: ✭ 5,489 (+831.92%)
Mutual labels:  artificial-intelligence, ai
Artificio
Deep Learning Computer Vision Algorithms for Real-World Use
Stars: ✭ 326 (-44.65%)
Mutual labels:  artificial-intelligence, ai
Spacy
💫 Industrial-strength Natural Language Processing (NLP) in Python
Stars: ✭ 21,978 (+3631.41%)
Mutual labels:  artificial-intelligence, ai
Caer
High-performance Vision library in Python. Scale your research, not boilerplate.
Stars: ✭ 452 (-23.26%)
Mutual labels:  artificial-intelligence, ai
Commandcenter
Starcraft AI Bot
Stars: ✭ 456 (-22.58%)
Mutual labels:  artificial-intelligence, ai
Autoscraper
A Smart, Automatic, Fast and Lightweight Web Scraper for Python
Stars: ✭ 4,077 (+592.19%)
Mutual labels:  artificial-intelligence, ai
Text Image Augmentation
Geometric Augmentation for Text Image
Stars: ✭ 333 (-43.46%)
Mutual labels:  recognition, detection
Rivescript Js
A RiveScript interpreter for JavaScript. RiveScript is a scripting language for chatterbots.
Stars: ✭ 350 (-40.58%)
Mutual labels:  artificial-intelligence, ai
Php Opencv Examples
Tutorial for computer vision and machine learning in PHP 7/8 by opencv (installation + examples + documentation)
Stars: ✭ 333 (-43.46%)
Mutual labels:  recognition, detection
Interpret
Fit interpretable models. Explain blackbox machine learning.
Stars: ✭ 4,352 (+638.88%)
Mutual labels:  artificial-intelligence, ai
Godot 2d Space Game
A 2D space exploration and mining game made with Godot and our AI framework
Stars: ✭ 462 (-21.56%)
Mutual labels:  artificial-intelligence, ai
Lightnet
🌓 Bringing pjreddie's DarkNet out of the shadows #yolo
Stars: ✭ 322 (-45.33%)
Mutual labels:  artificial-intelligence, ai
Maixpy scripts
micropython scripts for MaixPy
Stars: ✭ 326 (-44.65%)
Mutual labels:  ai, demo
Kglib
Grakn Knowledge Graph Library (ML R&D)
Stars: ✭ 405 (-31.24%)
Mutual labels:  artificial-intelligence, ai

本Demo 是为了在android上跑TensorFlow模型的 方便那些想把TensorFlow官网上的demo集成到自己项目里却又找不到头绪的人使用

正所谓前人栽树后人乘凉

https://github.com/tensorflow/tensorflow

https://github.com/ildoonet/tf-pose-estimation

Denis Tome, Chris Russell, Lourdes Agapito提出的Convolutional 3D Pose Estimation from a Single Image论文

在此特别感谢上述作者,喜欢原作的可以去使用原项目。同时欢迎大家下载体验本项目,如果使用过程中遇到什么问题,欢迎反馈。

如果需要自动打标请使用:https://github.com/yuxitong/AutoMarKingTensorFlowPython

另外特别感谢

[zyxcambridge](https://github.com/zyxcambridge) 童鞋
[manoshape](https://github.com/manoshape) 童鞋
[seriouslyhao](https://github.com/seriouslyhao) 童鞋
    	本项目模型由zyxcambridge、manoshape、seriouslyhao提供

本项目使用的是Camera2的api

过段时间我还会陆续公布一些更好玩的模型

如果想要体验本项目,请直接下载apk:点击下载

目前本Demo模型能识别出 抽烟 打电话 闭眼 睁眼

TensorFlowObjectDetectionAPIModel 为检测规则画框

TensorFlowImageClassifier2 为车道检测之后不规则绘制(因时间仓促 还没有进行绘图优化) 识别道路的测试方法请自行百度寻找图片或者视频都可以

TensorFlowImageClassifier3 是用来识别人体骨架的 这个模型是有特定输入和特定输出的 需要经过3层转换 才能使用 接下来准备上线道路障碍物识别... 最新版骨架识别目前支持区分各个身体部位具体情况请看注释

Camera2BasicFragment4 这是一个用检测来识别车道和前车 里面增加了点逻辑来判断是否是车道偏离或者前车过近 具体做法是 如果检测出线则判断斜率k = (y2-y1)/(x2-x1)然后设定一个固定斜率来判断是否是车道偏离 如果是检测出前面的车辆中心点在横屏8分之2到8分之6的范围内则判断中心点居上距离大于一定范围则算前车过近 或者如果车的高度大于一定级别则算前车过近

另外: 有人私下问我本项目在他们的手机上跑起来卡顿严重 这是算力的问题,目前tensorFlow在移动设备上貌似不支持GPU,而CPU的浮点运算速度比较慢导致的 推荐使用华为P10 或者 骁龙845 635之类的U来跑跑看 一般P10的话 1能一秒4帧 2能1秒8帧 3能一秒1帧 4能一秒6帧左右 当然以上数据仅供产考

目前人体骨架识别有了新突破:具体算法已经有了,但是暂时没工夫写。过段时间在上线。(把所有点连接成一个一个的人体)

具体算法如下:

WH19每一层的解释

         Nose = 0
         Neck = 1
         RShoulder = 2
         RElbow = 3
         RWrist = 4
         LShoulder = 5
         LElbow = 6
         LWrist = 7
         RHip = 8
         RKnee = 9
         RAnkle = 10
         LHip = 11
         LKnee = 12
         LAnkle = 13
         REye = 14
         LEye = 15
         REar = 16
         LEar = 17
         Background = 18

需要连线的身体部位,共19个连线

    CocoPairs = [
        (1, 2), (1, 5), (2, 3), (3, 4), (5, 6), (6, 7), (1, 8), (8, 9), (9, 10), (1, 11),
        (11, 12), (12, 13), (1, 0), (0, 14), (14, 16), (0, 15), (15, 17), (2, 16), (5, 17)
    ]   # = 19

对应每一个连线到 WH38 里找到两页取两组10个点

        CocoPairsNetwork = [
            (12, 13), (20, 21), (14, 15), (16, 17), (22, 23), (24, 25), (0, 1), (2, 3), (4, 5),
            (6, 7), (8, 9), (10, 11), (28, 29), (30, 31), (34, 35), (32, 33), (36, 37), (18, 19), (26, 27)
         ]  # = 19

从头开始按照顺序连接 到脖子 到右臂等等

    求出Dx = x1 - x2
        Dy = y1 - y2

计算两点间距离公式

    k = 根号下(Dx平方-Dy平方)

如果两点间距离小于0.0001 则分数为0 个数为0

计算 Vx = Dx / k Vy = Dy / k

连线的过程中每个线均分10个点 分出2个数组 分别是X数组 Y数组 每个数组都是10个数字 均分公式如下

    for(i = x1; i <= x2; i += dx/10)
            x[n] = i;
    Y同理

计算每个连线的分数: CocoPairs 与 CocoPairsNetwork 一一对应 每一组对应的CocoPairs连线方式可以在CocoPairsNetwork里取出2层 (如:1和2之间连线 可以在12层和13层里分别取出10个均等值) 并把均等值存到2个新的长度为10的组里 Px[10] 和 Py[10]

建立一个分数数组 = Px[10]乘Vx与Py[10]乘Vy对应相加

    分数>0.2的个数乘 >0.2的分数相加 得到最终分数,并且保留有多少个>0.2的
    之前两点间距离小于0.0001分数为0的个数也为0

如果个数小于5或者分数小于等于0则这个连线不要

从分数最高的开始遍历所有要的连线放到最终连线的集合(放的时候任何与这个连线集合的点重复都直接舍弃连线)

重复以上所有算法直到没点可连

然后拿出所有连线的集合 找出相同点的连线连接 就出现了人体骨架

image image image

image

Note that the project description data, including the texts, logos, images, and/or trademarks, for each open source project belongs to its rightful owner. If you wish to add or remove any projects, please contact us at [email protected].