PASCAL Dataset extend anotations

Part 2 & Alumni (2018)
1

From the website
http://host.robots.ox.ac.uk/pascal/VOC/voc2012/index.html#devkit

The training data provided consists of a set of images; each image has an annotation file giving a bounding box and object class label for each object in one of the twenty classes present in the image. Note that multiple objects from multiple classes may be present in the same image. Annotation was performed according to a set of guidelines distributed to all annotators.

Person: person
Animal: bird, cat, cow, dog, horse, sheep
Vehicle: aeroplane, bicycle, boat, bus, car, motorbike, train
Indoor: bottle, chair, dining table, potted plant, sofa, tv/monitor

Is there any way to extend this categories to say this is @jeremy or this is @gerardo instead of just a person?

2

The bounding box algorithm(localisation) will be the same and as for the other part, we should go for a multilabel(or a nested multiclass) classification pipeline.
I’m not sure… But, this is the first thing that popped up in my mind.

There is the possibility that the bounding box will occur between two people(from the lecture), and it would be hard to classify them afterwards.
Maybe in the upcoming lectures, we will find a way to predict better binding boxes.