In this paper we propose an automatic marine life monitoring system. First task in the monitoring process is to detect underwater moving objects as fishes. Second Task is to identify the species of the detected fish. Third task is to track the detected fish to avoid multiple counting and record their activities. Detection is performed using GMM based background subtraction method, classification is performed using Pyramid Histogram Of visual Words (PHOW) features with SVM classifier and finally identified fishes are tracked using “Kalman Filter”. This experiment is performed using data-set from the CLEF 2015. The proposed system can detect and track fishes with 48.94 percent accuracy in videos, and it can identify fishes in high resolution still image with 91.7 percent accuracy where as in the low quality video fishes are detected with 40.1 percent accuracy.

Feature selection and discretization have been considered to be an important research topic in the field of pattern recognition and data mining. However, addressing both these issues at a time is rarely discussed in the existing research. In this paper, these issues have been addressed by developing a heuristic namely discretization and selection of features based on mutual information (DSM). Experimental results on 15 datasets show that the proposed DSM outperforms a number of state-of-the-art feature selection or discretization algorithms. On average, its accuracy surpasses that of the best performing state-of-the-art algorithms by 5% using Support Vector Machine. Moreover, for datasets with a large number of features, it shows promising accuracies even with far less number of features than the other competing algorithms.

This paper introduces a new splitting criterion called Inter-node Hellinger Distance (iHD) and a weighted version of it (iHDw) for constructing decision trees. iHD measures the distance between the parent and each of the child nodes in a split using Hellinger distance. We prove that this ensures the mutual exclusiveness between the child nodes. The weight term in iHDw is concerned with the purity of individual child node considering the class imbalance problem. The combination of the distance and weight term in iHDw thus favors a partition where child nodes are purer and mutually exclusive, and skew insensitive. We perform an experiment over twenty balanced and twenty imbalanced datasets. The results show that decision trees based on iHD win against six other state-of-the-art methods on at least 14 balanced and 10 imbalanced datasets. We also observe that adding the weight to iHD improves the performance of decision trees on imbalanced datasets. Moreover, according to the result of the Friedman test, this improvement is statistically significant compared to other methods.