Incremental concept learning with few training examples and hierarchical classification

Incremental concept learning with few training examples and hierarchical classification

Bouma, H.
Eendebak, P.T.
Schutte, K.
Azzopardi, G.
Burghouts, G.J.

Burgess, D. (editor)

2015

Object recognition and localization are important to automatically interpret video and allow better querying on its content. We propose a method for object localization that learns incrementally and addresses four key aspects. Firstly, we show that for certain applications, recognition is feasible with only a few training samples. Secondly, we show that novel objects can be added incrementally without retraining existing objects, which is important for fast interaction. Thirdly, we show that an unbalanced number of positive training samples leads to biased classifier scores that can be corrected by modifying weights. Fourthly, we show that the detector performance can deteriorate due to hard-negative mining for similar or closely related classes (e.g., for Barbie and dress, because the doll is wearing a dress). This can be solved by our hierarchical classification. We introduce a new dataset, which we call TOSO, and use it to demonstrate the effectiveness of the proposed method for the localization and recognition of multiple objects in images © (2015) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE).

Observation, Weapon & Protection Systems ICT
II - Intelligent Imaging BIS - Business Information Services
TS - Technical Sciences
Defence Research
Image processing
Defence, Safety and Security
Image data
Object recognition
Deep learning
Online learning
Hierarchical clustering.

http://resolver.tudelft.nl/uuid:167c0171-dc7d-47e7-be9b-2ae290bacdcd

https://doi.org/10.1117/12.2194438

528828

SPIE, Bellingham, WA

Optics and Photonics for Counterterrorism, Crime Fighting, and Defence XI; and Optical Materials and Biomaterials in Security and Defence Systems Technology XII, 21 September 2015, Toulouse France

Proceedings of SPIE

6520E

conference paper