Object Detection and Recognition by Image Parsing using Matlab Wavelet Technique
Shrikant N. Sarda, V. T. Ingole, Virupakshappa Patil · 2013
These systems propose a general framework for parsing images into regions and objects. In this framework, the detection and recognition of objects proceed simultaneously with image segmentation in a competitive and cooperative manner. This approach illustrates natural images of complex city scenes where the objects of primary interest are faces and text. This method makes use of bottom,up proposals combined with top,down generati ve models algorithm which is guaranteed to converge to the optimal estimate asymptotically. More precisely, it is define generative models for faces, text, and generic regions- e.g. shading, texture, and clutter. These models are activated by bottom,up proposals. This system illustrate the advantages and importance of combining bottom,up and Top,down models and of performing segmentation and object detection/recognition simultaneously.