Adaptable Models and Semantic Filtering for Object Recognition in Street Images
- When?
- Thursday 1 October 2009, 11:00 to 12:00
- Where?
- 39 BB 02
- Open to:
- Staff, Students
- Speaker:
- Qin Ge
The need for a generic and adaptable object detection and recognition method in static images is becoming a necessity today, given the rapid development of the internet and multimedia databases in general. Comparing with human vision, the computer vision is out-performed in terms of efficiency, accuracy and depth of understanding, as the computerised recognition is achieved at contextual level. In order to achieve recognition at semantic level, computer vision systems must not only be capable of recognising objects, regardless of the changes in appearance, location, and action, but also be able to interpret abstract non-observable concepts.
This work reviews the-state-of-the-art techniques in object recognition and proposes a supervised learning method based on adaptable models for detecting thematic categories of objects. Investigation is performed extracting visual characteristics, which can be used to distinguish the object; and analysis is performed examining the semantic relations, which are embedded within the object’s inner-parts and its surroundings. Such visual characteristics together with the semantic relations can be considered as a distinctive way for object recognition. By constantly identifying and capturing the associations between individual components forming that object, the set of unique visual characteristics of one object can be revealed. This method goes beyond classical image indexing and retrieval techniques, based on visual features only, and introduce semantics as a way of recognising and retrieving objects. The ultimate goal of this work is to develop a knowledge guided object recognition method, which concentrates on extracting hidden semantic information to support the recognition.
The proposed method is able to achieve objects recognition in real life images with reasonable high accuracy (83-87% for foreground objects recognition and 87-94% for background objects recognition). More importantly, semantic reasoning is introduced to guide and assist the recognition process. Instead of relying on purely visual similarity measure, objects are recognised based on their key components observed together with the understanding about their relations with surrounding environment.