Ayyalasomayajula, PradyumnaGrassi Pauletti, SaraFarine, Pierre-André2011-09-092011-09-092011-09-09201110.1109/PRIME.2011.5966243https://infoscience.epfl.ch/handle/20.500.14299/70844We propose a low complexity image recognition algorithm based on Content Based Image Retrieval (CBIR) suitable for handheld applications. The target application is an Alternative and Augmentative Communication (AAC) device used in speech rehabilitation and education. The device recognizes images (pictograms and pictures) and plays a sound message associated with the recognized image. Experimental validation of the proposed algorithm using MATLAB and its DSP implementation is presented.Image retrievalContent based image retrievalAlternative and Augmentative CommunicationHandheld image recognition systemtext::conference output::conference proceedings::conference paper