SDD 0.2.2.1

lightweight video detection

SocialDistance

Keep safe social distance is considered as an effective way of avoiding spreading of coronavirus. Our SocialDistance module SDD is a lightweight package which provides an implementation of utlizing deep learning models for monitoring safe social distance.

Demo

Watch the demo video

Dataset

We use the video clip collected from OXFORD TOWN CENTRE dataset and made the above demo video.

Supported Models

We have tested our model using Faster-RCNN, YOLO-v3 and SSD, based on the performance of each model, we have chosen YOLO-v3 as our default model

All our models are pretrained models from Gluno CV Tookit

Installation

You may be able to obtain the latest version our model from:

pip install SDD

How to Use

After Successfully installed SocialDistance, you can use it for detection by:

import SDD
from SDD.utils.Run import Detect
detect = Detect(pretrained_models = 'yolo_v3')
detector = detect(save_path = output_path_image, video = False, device = mx.cpu())
_ = detector(file_path)

Parameters

• pretrained_models: str, Currently, we provided two pretrained models: 'yolo_v3' and 'ssd'
• save_path: str, Path where you want to save the output video/ images
• video: boolean, If your input is a video, set this parameter as True, if your input is a set of images, set this parameter as False
• device: mx.cpu() or mx.gpu()
• file_path: str, Input path of your video or image folder

Reference

1. Landing AI 16 April 2020, Landing AI Creates an AI Tool to Help Customers Monitor Social Distancing in the Workplace, accessed 19 April 2020, https://landing.ai/landing-ai-creates-an-ai-tool-to-help-customers-monitor-social-distancing-in-the-workplace/