Malaria microscopy, microscopic examination of stained blood slides to detect\nparasite Plasmodium, is considered to be a gold-standard for detecting\nlife-threatening disease malaria. Detecting the plasmodium parasite requires a\nskilled examiner and may take up to 10 to 15 minutes to completely go through\nthe whole slide. Due to a lack of skilled medical professionals in the\nunderdeveloped or resource deficient regions, many cases go misdiagnosed;\nresulting in unavoidable complications and/or undue medication. We propose to\ncomplement the medical professionals by creating a deep learning-based method\nto automatically detect (localize) the plasmodium parasites in the photograph\nof stained film. To handle the unbalanced nature of the dataset, we adopt a\ntwo-stage approach. Where the first stage is trained to detect blood cells and\nclassify them into just healthy or infected. The second stage is trained to\nclassify each detected cell further into the life-cycle stage. To facilitate\nthe research in machine learning-based malaria microscopy, we introduce a new\nlarge scale microscopic image malaria dataset. Thirty-eight thousand cells are\ntagged from the 345 microscopic images of different Giemsa-stained slides of\nblood samples. Extensive experimentation is performed using different CNN\nbackbones including VGG, DenseNet, and ResNet on this dataset. Our experiments\nand analysis reveal that the two-stage approach works better than the one-stage\napproach for malaria detection. To ensure the usability of our approach, we\nhave also developed a mobile app that will be used by local hospitals for\ninvestigation and educational purposes. The dataset, its annotations, and\nimplementation codes will be released upon publication of the paper.\n