A Fruit Recognition Method via Image Conversion Optimized through Evolution Strategy

This research is to propose a fast and highly accurate object recognition method especially for fruit recognition applications to be used in a mobile environment. Conventional techniques are based on one or more of the basic features that characterize an object: color, shape, texture and intensity, causing performance or accuracy limitations in a mobile environment. Thus, this paper presents a combined approach that transforms basic features into their associated code fields to generate an object code that could be used as a search key in a feature database. Parameters used in the experiment have been optimized by using Evolution Strategies and an increase in accuracy by up to 10% has been achieved. A fruit database consisting of 36 different classes of fruits and 1108 fruit images overall has been used to obtain the experimental results. The results show an average accuracy of more than 98% and performance increase compared to different approaches on fruit image recognition.