Skip to content

Toy project with Detectron2 to do object recognition on a single class: cars.

Notifications You must be signed in to change notification settings

Venn1998/ObjectRecognition_Detectron2

Repository files navigation

OBJECT DETECTION _ Lorenzo Venieri

The source code can be found in the notebook: “Cars_Detection.ipynb”

The project was developed on Google Colab.

To ease the data preparation I’ve moved the “annotations_sample.json” file out of the “cars” folder

The output with the car boxes in (Cx,Cy,W,H) format is in the file: “car_bboxes_RightFormat.csv”

The output in format (x1,y1,x2,y2) is in the file: “car_bounding_boxes.csv” , I’ve used it in the notebook because this format is easier to work with when using Detectron2.

The bounding boxes of the cars detected as red is in the file “red_cars_bboxes.csv”, boxes are in XYXY_ABS format.

To train and validate the model using Detectron2 it was useful to generate two json files “cars_annotations_train.json” and “cars_annotations_val.json” with the annotations for the images in COCO format.

About

Toy project with Detectron2 to do object recognition on a single class: cars.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published