Tuna classification using super learner ensemble of region-based CNN-grouped 2D-LBP models

Tuna is superior among the marine fishes that are exported in the forms of raw fish and processed food.Separation of Tuna into their species is done in industries manually,and the process is tiresome.This work proposes an automated system for classifying Tuna spe-cies based on their images.An ensemble of region-based deep neural networks is used.A sub region contrast stretching operation is applied to enhance the images.Each fish image is then divided into three regions and is augmented before giving as input to pre-trained convolutional neural networks(CNN).After fine-tuning the models,the output from the last convolutional layer is given to a grouped 2D-local binary pattern descriptor(G2DLBP).Statistical features from the descriptor are applied to different classifiers,and the best clas-sifier for each image region model is identified.Different ensemble methods are subse-quently used to combine the three CNN-G2DLBP models.Among the ensemble techniques,super learner ensemble method with random forest(RF)classifier using 5-fold cross-validation shows the highest classification accuracy of 97.32%.The perfor-mance of different ensemble methods is analyzed in terms of accuracy,precision,recall,and f-score.The proposed system shows an accuracy of 93.91% when evaluated with an independent test dataset.An ensemble of region-based CNN with textural features from G2DLBP is applied for the first time for fish classification.

EXTRA-SEASONAL PREDICTIONS OF SUMMER RAINFALL IN CHINA AND ENSO IN 2001 BY CLIMATE MODELS

China is a monsoon country.The most rainfalls in China concentrate on the summer seasons. More frequent floods or droughts occur in some parts of China.Therefore,the prediction of summer rainfall in China is a significant issue.As we know,the obvious impacts of the sea surface temperature anomalies(SSTA)on the summer rainfall over China have been noticed.The predictions of the SSTA have been involved in the research. The key project on short-term climate modeling prediction system has been finished in 2000. The system included an atmospheric general circulation model named AGCM95,a coupled atmospheric-oceanic general circulation model named AOGCM95,a regional climate model over China named RegCM95,a high-resolution Indian-Pacific OGCM named IPOGCM95,and a simplified atmosphere-ocean dynamic model system named SAOMS95.They became the operational prediction models of National Climate Center(NCC). Extra-seasonal predictions in 2001 have been conducted by several climate models,which were the AGCM95,AOGCM95,RegCM95,IPOGCM95,AIPOGCM95,OSU/NCC,SAOMS95,IAP APOGCM and CAMS/ZS.All of those models predicted the summer precipitation over China and/ or the annual SSTA over the tropical Pacific Ocean in the Modeling Prediction Workshop held in March 2001. The assessments have shown that the most models predicted the distributions of main rain belt over Huanan and parts of Jiangnan and droughts over Huabei-Hetao and Huaihe River Valley reasonably.The most models predicted successfully that a weaker cold phase of the SSTA over the central and eastern tropical Pacific Ocean would continue in 2001. The evaluations of extra-seasonal predictions have also indicated that the models had a certain capability of predicting the SSTA over the tropical Pacific Ocean and the summer rainfall over China.The assessment also showed that multi-model ensemble(super ensembles)predictions provided the better forecasts for both SSTA and summer rainfall in 2001,compared with the single model. It is a preliminary assessment for the extra-seasonal predictions by the climate models.The further investigations will be carried out.The model system should be developed and improved.