https://github.com/kleinyuan/tf-object-detection

Simpler app for tensorflow object detection API
https://github.com/kleinyuan/tf-object-detection

faster-rcnn faster-rcnn-inception-resnet faster-rcnn-resnet mobilenet object-detection rfcn-resnet ssd ssd-inceptionv2 ssd-mobilenet tensorflow tensorflow-examples tensorflow-models

Last synced: 2 months ago
JSON representation

Simpler app for tensorflow object detection API

• Host: GitHub
• URL: https://github.com/kleinyuan/tf-object-detection
• Owner: KleinYuan
• License: mit
• Created: 2017-09-13T07:44:44.000Z (almost 8 years ago)
• Default Branch: master
• Last Pushed: 2018-06-09T06:41:20.000Z (about 7 years ago)
• Last Synced: 2025-04-07T02:03:05.216Z (3 months ago)
• Topics: faster-rcnn, faster-rcnn-inception-resnet, faster-rcnn-resnet, mobilenet, object-detection, rfcn-resnet, ssd, ssd-inceptionv2, ssd-mobilenet, tensorflow, tensorflow-examples, tensorflow-models
• Language: Python
• Size: 4.57 MB
• Stars: 93
• Watchers: 6
• Forks: 37
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Scope

This repo is aiming to provide production ready **2D object detection** code basics.

It's based on official tensorflow API jupyter notebook but I will gradually add more popular models such as `yolo` series.

### More

If you are interested in **3D object detection**, visit this [repo](https://github.com/KleinYuan/tf-3d-object-detection).

If you are interested in **Segmentation**, visit this [repo](https://github.com/KleinYuan/tf-segmentation).

# Introduction

This is a repo for implementing object detection with pre-trained models (as shown below) on tensorflow.

| Model name  | Speed | COCO mAP | Outputs |

| ------------ | :--------------: | :--------------: | :-------------: |

| [ssd_mobilenet_v1_coco](http://download.tensorflow.org/models/object_detection/ssd_mobilenet_v1_coco_11_06_2017.tar.gz) | fast | 21 | Boxes |

| [ssd_inception_v2_coco](http://download.tensorflow.org/models/object_detection/ssd_inception_v2_coco_11_06_2017.tar.gz) | fast | 24 | Boxes |

| [rfcn_resnet101_coco](http://download.tensorflow.org/models/object_detection/rfcn_resnet101_coco_11_06_2017.tar.gz)  | medium | 30 | Boxes |

| [faster_rcnn_resnet101_coco](http://download.tensorflow.org/models/object_detection/faster_rcnn_resnet101_coco_11_06_2017.tar.gz) | medium | 32 | Boxes |

| [faster_rcnn_inception_resnet_v2_atrous_coco](http://download.tensorflow.org/models/object_detection/faster_rcnn_inception_resnet_v2_atrous_coco_11_06_2017.tar.gz) | slow | 37 | Boxes |

Dependencies:

- [X] Tensorflow >= 1.2.0

- [X] OpenCV

# Run Demo

```

# Clone this repo

git clone https://github.com/KleinYuan/tf-object-detection.git

# Setting up

cd tf-object-detection

bash setup.sh

# Run demo

python app.py

```

![res](https://user-images.githubusercontent.com/8921629/32482793-24968e20-c34e-11e7-9810-4aef685d067f.jpg)

# Image Classifications

I also put an image classifications inference app (VGG16) here.

```

# Assuming you already run setup.sh, which will download vgg16.np

python app_ic.py

```

# Networks

| Model name  | Architecture|

| ------------ | :--------------: |

| [AlextNet](https://papers.nips.cc/paper/4824-imagenet-classification-with-deep-convolutional-neural-networks) | ![AlexNet](https://kratzert.github.io/images/finetune_alexnet/alexnet.png)|

| [Vgg 16](https://arxiv.org/abs/1409.1556) | ![VGG16](https://www.cs.toronto.edu/~frossard/post/vgg16/vgg16.png)|

| [SSD](https://arxiv.org/abs/1512.02325) | ![SSD](http://joshua881228.webfactional.com/media/uploads/ReadingNote/arXiv_SSD/SSD.png)|

| [ResNet](http://arxiv.org/abs/1512.03385)|![Resnet](https://image.slidesharecdn.com/lenettoresnet-170509055515/95/lenet-to-resnet-17-638.jpg)|

| [MobileNet](https://arxiv.org/abs/1704.04861)|![MobileNet](http://machinethink.net/images/mobilenets/[email protected]) |

| [Faster R-CNN](https://arxiv.org/abs/1506.01497) | ![fasterrcnn](https://raw.githubusercontent.com/sunshineatnoon/Paper-Collection/master/images/faster-rcnn.png)|