EVALUATION OF MORPHOLOGICAL RESPONSES IN PARMOTREMA TINCTORUM LICHEN USING BACK PROPAGATION NEURAL NETWORKS COLLECTED FROM SERVARAYAN HILLS OF TAMILNADU, INDIA
Kanmani P, Rajiv Kannan A · International Research Journal of Pharmacy · 2017
Lichens are combination of both fungi and algae which is used as one of the key pharmaceutical ingredients in medicinal industry.The traditional way of identification of lichens requires skilled person and also time consuming process for performing the manual based colour test.The present study was planned to develop a reliable technique to identify the lichen species through image processing and machine learning process.Lichen sample was collected and image acquisition was done.Lichen images were preprocessed and to extract the features of lichens by color co-occurrence method.Segmentation of lichen images were performed using K means clustering algorithm.Back propagation neural network was used to automate the identification process.30 images were taken for training and 40 images were taken for testing.Automation by Neural network revealed that 87% of accuracy in recognising the lichen images.A comparison between the manual and the proposed system identification is evaluated.The results showed that the parameters time, cost had reduced and the efficiency and performance were increased in the proposed system.Further, the automation techniques of the present study will enable to identify the lichen species in easiest approach which helps to predict the important lichen species for pharma industry.