Overview of the ImageCLEF 2017 Tuberculosis Task - Predicting Tuberculosis Type and Drug Resistances

ImageCLEF is the image retrieval task of the Conference and Labs of the Evaluation Forum (CLEF). ImageCLEF has historically focused on the multimodal and language-independent retrieval of images. Many tasks are related to image classification and the annotation of image data as well as the retrieval of images. The tuberculosis task was held for the first time in 2017 and had a very encouraging participation with 9 groups submitting results to these very challenging tasks. Two tasks were proposed around tuberculosis: (1) the classification of the cases into five types of tuberculosis and (2) the detection of drug resistances among tuberculosis cases. Many different techniques were used by the participants ranging from Deep Learning to graph-based approaches and best results were obtained by a large variety of approaches. The prediction of tuberculosis types had relatively good performance but the detection of drug resistances remained a very difficult task. More research into this seems necessary.