Learning complex cell features with cooperating pooling operation for object recognition

PROJECT TITLE :

Learning complex cell features with cooperating pooling operation for object recognition

ABSTRACT :

A straightforward biologically inspired feature extraction algorithm is proposed for object recognition. 1st, a group of statistical topographic filters modelling the properties of complicated cells in an exceedingly primary visual cortex (V1) are learned primarily based on enhanced freelance subspace analysis (EISA), and locally invariant feature maps are extracted by convolving the filters with each image. Then, the cooperating cortical pooling operations which mix the energy model and the MAX-like model are used to increase the section and shift invariance of the filter response. Experimental results on the MNIST dataset and the Caltech101 dataset demonstrate that the algorithm is economical and achieves high recognition accuracy.

• Programmable and implantable neurostimulator with novel stimulus waveforms for rat models
• Half-cycle modulation for VCSEL based 6-Gbaud 4-QAM transmission over 1 km multimode fibre link