Closed Niranjankumar-c closed 5 years ago
You can do it by masking the output.
model.predict
function of detection model returns bboxes, labels, scores
, labels
have class information.
you can filter bounding boxes with something like human_bboxes = bboxes[labels == label_names.index('human')]
Can i use the ChainerCV to detect only the objects that belong to single class, like i want to detect only persons face and draw a bounding box around that face using frcnn or ssd? is it possible?.