Parasitic infections are one of the main causes of infectious diseases in animals. Depending on the shape of the parasite, an exact parasite causing infection can be identified and suitable deworming agent can be suggested. Microscopic images are typically used to make a diagnosis, with error rates ranging from modest to substantial. Computational image analysis has been used to solve the problem. A total of 650 images of the seven most common species of parasitic ova are taken, namely Amphistome, Ascaris Egg, B. Coli, Moniezia Ova, Schistosoma Spindale, Strongyle, and Trichuris Egg. In this paper, input microscopic images are segmented to detect parasites in the image, it is accomplished by semantic segmentation using DeepLabv3+ architecture and Inception V3 is used as a backbone for prediction purpose. The model has been trained for 70 epochs considering the batch size as 4 and the number of classes as 8 (i.e. including the background as a class). Then the model is evaluated, the images are given for testing. Out of 650 images, 500 images are utilised for training and 150 for testing and validation. The obtained accuracy is 99.6%, precision 99.7%, Recall 99.9%, F1 score 99.8% and Jaccard 99.6%. The proposed technique is used for faster and accurate diagnosis of parasitic infection and expected to replace the problem of lack of skilled experts to some extent.