- Free software: MIT license
- Documentation: https://object-detection.readthedocs.io.
In the past, the tasks of Computer Vision were to classify MNIST numbers or to distinguish images of dogs from images of cats. Now convolutional neural networks can perform much more complex image processing tasks. One such task is object recognition. In particular, YOLO and SSD neural net- works can recognize instances of different objects and highlight their bounding boxes in photos, even if there are several such instances, they have different sizes and partially overlap. Our purpose is to test these neural networks on our own dataset with images of books. Our goal is to recognize books and build bounding boxes for them using YOLO and SSD models.
This is basically a set of pictures of my books in the dormitory! and Books are labeled by the author/publisher. We have 8 classes + background. roboflow books-umdmd dataset
our notebooks will help you out in here.
- Aarohi for her good video tutorials on Object Detection models.
- ultralytics for his good of YOLOv5integration with roboflow.
- roboflow-ai for the good dataset service and implementation of RCNN.