Camouflaged objects are difficult to detect, for both humans and artificial intelligence (AI). But now an AI has been trained to parse objects from their backgrounds.
As well as the obvious military setting, this could have a variety of applications such as:
- Search-and-rescue work;
- Detecting agricultural pests; and
- Medical imaging.
Detecting camouflaged objects requires visual perception and knowledge. Until now, many AIs have struggled with this task because their algorithms rely on visual cues, such as differences in colour or easily recognisable shapes, to identify objects.
To improve on this, Fan et al. (2020) collated a data set of 10,000 photographs to train an AI. The data set includes 5,066 images of camouflaged objects, which they divided into 78 categories, such as “amphibian”, “aquatic” and “flying”. Although databases of camouflaged objects already exist, this data set is the largest. The photographs included:
- Naturally camouflaged animals (e.g. fish and insects); andA
- Artificial camouflage (e.g. soldiers in uniform).
The team manually labelled each image of a camouflaged object to highlight characteristics such as its shape or whether it was partially obstructed by its surrounding environment. They then developed an AI called SINet and trained it on images from the data set.
The researchers compared SINet to 12 existing algorithms built to detect generic objects. They tested all 13 algorithms using three existing data sets of camouflaged objects. SINet did better than the other 12 at isolating camouflaged objects and identifying their correct shape and nature in both the existing and the training data sets.
The researchers hope the data set and algorithm can improve AI’s ability to recognise camouflaged objects.
Fan, D-P., Ji, G-P., Sun, G., Cheng, M-M., Shen, J. & Shao, L. (2020) Camoflaged Object Detection. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp.2777-2787.