Abstract: Objective To facilitate the identification of parasite eggs using computer technology, establish the automation-based applications, and propose an algorithm for egg classification. Methods Eggs of 11 parasites, Clonorchis sinensis, Taenia solium, Enterobius vermicularis, Ascaris lumbricoides, Trichuris trichiura, Spirometra mansoni, Diphyllobothrium latum, Ancylostoma duodenale, Schistosoma japonicum, Paragonimus westermani and Fasciolopsis buski, were selected and divided into two groups, the training group and the testing group, and were microphotographed. The eigenvalue was extracted using the VC++-based method. The eigenvalue database was constructed, and the training data set was tested with a variety of classification algorithms. The classifier was constructed using algorithm with the highest efficiency and an identification method was established by multi-feature fusion. Results After removal of images with invalid values, the training group received 19 844 egg images, and the testing group, 3 721 images. Based on the 14 eigenvalues, there were significant differences in the size and color among the eggs of 11 parasite species. For example, the length, width, area and brightness of the smallest parasite egg of Clonorchis sinensis were 292.24 μm, 192.64 μm, 43 416.61 μm2 and 53.84, respectively, while those of the largest parasite egg of Fasciolopsis buski were 945.31 μm, 610.88 μm, 536 002.60 μm2 and 100.54, respectively. When using dynamic weights to construct the classifier, the discrimination rate on the training data set was 88.89%（17 641/19 844）, and that on the verification data set was 91.83%（3 004/3 271）, with an average modeling time of 0.01 s. Conclusion The algorithm for egg classification has been established, which pravides a basis for further study on its feasibility．

Key words: Parasite eggs, Multiple features, Adaptive fusion, Classification algorithm