This repository containes the code for the Logo Detection in Video Project
In this demo OCI AI Vision Service is used to detect the presence of a logo inside a Video
- OpenCV (cv2)
- OCI AI Vision Service, for Object Detection
- OCI Data Labelling Service (DLS)
Around 140 images, stored in a bucket in the Object Storage, have been labelled using OCI DLS
OCI AI Vision has been used. A model has been trained on the labelled dataset for 8 hrs.
Metric. | Value |
---|---|
mAP@0.5 | 0.9038 |
Precision | 0.9886 |
Recall | 0.8978 |
The main NoteBook for the demo is here
To simplify the code using OCI AI Vision SDK, a Python class has beeen provide. See: oci_vision_utilities