Social Distancing

Social distancing implies that people should physically distance themselves from one another, reducing close contact, and thereby reducing the spread of a contagious disease

Final prototype

Sample street	Train Station

Structure of the project

• frames
  • sample1/ --> folder containing the frames of the video sample1
• video
  • sample1.mp4 --> input video1
  • sample1.txt --> information for video1
• models.py --> definition of object detection models and method for calculating social distance.
• trace_roi.py --> definition of the interface in open-cv to calculate the ROI
• utils.py --> set of methods useful for the video process and for calculating the transformed points
• Predict_SocialDistancing.ipynb --> notebook to launch on the cloud (colab) to use the models on the gpu.

Region of interest definition

Using first 4 points or coordinates for perspective transformation. The region marked by these 4 points are considered ROI. This polygon shaped ROI is then warped into a rectangle which becomes the bird eye view. The next 3 points are for horizontal and vertical unit length (in this case 180 cm)

Pedestrian Detection Model

To achieve object detection of pedestrians on the scene we used various models. State-of-the-art object detectors use deep learning approaches, which are usually divided into two categories. The first one is called two-stage detectors, mostly based on R-CNN, which starts with region proposals and then performs the classification and bounding box regression. The second one is called one-stage detectors (for example YOLO, SSD, RetinaNet and EfficientDet). We have used the pre-trained of these models which perform very well. To use them we used 2 very popular computer vision frameworks on the market: Detectron2 for the Faster R-CNN model and ImageAI for the YOLOv3 and YOLOtiny models.

Detectron2	ImageAI

At runtime the user can select the model with which to perform the object detection. YOLOv3 allows fast execution with good accuracy while Faster R-CNN provides better accuracy but higher computation times. The models are defined as follows:

def YoloV3_model(yolov3_model_path):
    '''
    Method that creates a YoloV3 model, using config from ImageAi core library.
    :param yolov3_model_path: the path of the config file
    :return: the YoloV3 model used to predict and the custom objects ( only poeple) to pass to the model
    during prediction.
    '''
    detector = ObjectDetection()
    detector.setModelTypeAsYOLOv3()  # Se vuoi usare yolo tiny cambia il set model
    detector.setModelPath(yolov3_model_path)
    custom_objects = detector.CustomObjects(person=True)
    detector.loadModel()
    return detector, custom_objects
    
def faster_RCNN_model():
    '''
    Method that creates a fasterRCNN model, using config and pretrained weights
    from detectron2 core library
    :return: the fasterRCNN model used to predict
    '''
    cfg = get_cfg()
    cfg.MODEL.DEVICE = 'cuda'

    # add project-specific config (e.g., TensorMask) here if you're not running a model in detectron2's core library
    cfg.merge_from_file(model_zoo.get_config_file("COCO-Detection/faster_rcnn_R_50_C4_3x.yaml"))
    cfg.MODEL.ROI_HEADS.SCORE_THRESH_TEST = 0.94  # set threshold for this model

    # Find a model from detectron2's model zoo. You can use the https://dl.fbaipublicfiles... url as well
    cfg.MODEL.WEIGHTS = model_zoo.get_checkpoint_url("COCO-Detection/faster_rcnn_R_50_C4_3x.yaml")
    predictor = DefaultPredictor(cfg)
    return predictor

Authors

• Prandini Manuel

• Bianca Francesco

• Berti Andrea

References

[1]. Dongfang Yang, Ekim Yurtsever, Vishnu Renganathan, Keith A. Redmill, Ümit Özgüner, “A Vision-based Social Distancing and Critical Density Detection System for COVID-19”, 2020.

[2]. Shaoqing Ren, Kaiming He, Ross Girshick, and Jian Sun, “Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks”, 2016.

[3]. Joseph Redmon, Ali Farhadi, “YOLOv3: An Incremental Improvement”.

[4]. Pranav Adarsh, Pratibha Rathi, Manoj Kumar, “YOLO v3-Tiny: Object Detection and Recognition using one stage improved model”, 2020 6th International Conference on Advanced Computing & Communication Systems (ICACCS).

[5]. Yuxin Wu and Alexander Kirillov and Francisco Massa and Wan-Yen Lo and Ross Girshick, "Detectron2", 2019, https://github.com/facebookresearch/detectron2

[6]. Moses and John Olafenwa, "ImageAI, an open source python library built to empower developers to build applications and systems with self-contained Computer Vision capabilities", mar 2018, https://github.com/OlafenwaMoses/ImageAI