Breast cancer is the most common malignancy in women. Mammographic findings\nsuch as microcalcifications and masses, as well as morphologic features of\nmasses in sonographic scans, are the main diagnostic targets for tumor\ndetection. However, improved specificity of these imaging modalities is\nrequired. A leading alternative target is neoangiogenesis. When pathological,\nit contributes to the development of numerous types of tumors, and the\nformation of metastases. Hence, demonstrating neoangiogenesis by visualization\nof the microvasculature may be of great importance. Super resolution ultrasound\nlocalization microscopy enables imaging of the microvasculature at the\ncapillary level. Yet, challenges such as long reconstruction time, dependency\non prior knowledge of the system Point Spread Function (PSF), and separability\nof the Ultrasound Contrast Agents (UCAs), need to be addressed for translation\nof super-resolution US into the clinic. In this work we use a deep neural\nnetwork architecture that makes effective use of signal structure to address\nthese challenges. We present in vivo human results of three different breast\nlesions acquired with a clinical US scanner. By leveraging our trained network,\nthe microvasculature structure is recovered in a short time, without prior PSF\nknowledge, and without requiring separability of the UCAs. Each of the\nrecoveries exhibits a different structure that corresponds with the known\nhistological structure. This study demonstrates the feasibility of in vivo\nhuman super resolution, based on a clinical scanner, to increase US specificity\nfor different breast lesions and promotes the use of US in the diagnosis of\nbreast pathologies.\n