Tensorflow 2 Object Detection API Demo

In this repository you can find some examples on how to use the Tensorflow OD API with Tensorflow 2. Referenced from Tensorflow Object Detection with Tensorflow 2 (Gilbert Tanner)

This TF2 OD example code was used for the demo in my talk, UP ACM Fest 2021 "AI Hackathon: Building Your First Prototype"

Accompanying slides can be found here: Google Slides Link

Recorded video can be viewed here (13:55 mark) Building AI Prototypes Workshop

I. Project Setup

1. Creating python virtual env

$ python3 -m venv ~/venvs/[name_of_env]

e.g.
$ python3 -m venv ~/venvs/tf-od

2. Activating python virtual env

$ source ~/venvs/[name_of_env]/bin/activate

e.g.
$ source ~/venvs/tf-od/bin/activate

3. Install the project's requirements

$ pip3 install -r requirements.txt

II. Download Pre-trained Model

1. Go to TF Object Detection Model Zoo

2. Choose one of the models and download

3. Extract the downloaded model

**For the example output, efficientdet_d2_coco17_tpu-32 was used

III. Video Inference

1. Upload your video in the vids folder

2. Run the following command but replace the parameters as needed

python3 detect_from_video.py --model=efficientdet_d2_coco17_tpu-32/ \
                                --video_path=vids/smarthome.mp4 \
                                --labelmap=mscoco_label_map.pbtxt

3. Your video output should be inside the outputs folder