Platelets are functionally versatile blood cells involved in thrombosis, hemostasis, atherosclerosis, and immune response. Platelet interaction with the immediate microenvironment in blood, vasculature, and tissues alters platelet morphology. The quantification of platelet morphodynamics by geometrical parameters (morphometry) can provide important insights into how platelets sense and respond to stimulatory cues in their vicinity. However, the extraction of platelet shapes from phase contrast microscopy images by conventional image processing is difficult. Here, we used a convolutional neural network (CNN) to develop a deep-learning-based approach for the unbiased extraction of information on platelet morphodynamics by phase contrast microscopy. We then investigated the effect of normal and oxidized low-density lipoproteins (LDL, oxLDL) on platelet morphodynamics, spreading, and haptotactic migration. Exposure of platelets to oxLDL led to a decreased spreading area and rate on fibrinogen, accompanied by increased formation of filopodia and impaired formation of lamellipodia. Haptotactic platelet migration was affected by both LDL and oxLDL in terms of decreased migration velocity and reduced directional persistence. Our results demonstrate the use of deep learning in investigating platelet morphodynamics and reveal differential effects of LDL and oxLDL on platelet morphology and platelet–matrix interaction.