Accurate　analysis　of　plaque　types　in　blood　vessels　can　not　only　prevent　the　occurrence　of　cardiovascular　diseases,　but　also　provide　a　reliable　basis　for　the　treatment　of　diseases.　Therefore,　in　view　of　the　low　accuracy　of　plaque　recognition　methods　in　intravascular　ultrasound　images　and　the　need　to　extract　features　manually,　this　paper　proposes　a　method　to　automatically　extract　image　features　and　recognize　plaques.　The　results　show　that　on　the　same　dataset,　the　accuracy　rates　of　the　Support　Vector　Machine　(SVM)　model　and　our　method　are　75.3%　and　87.1%,　respectively,　and　our　method　achieves　higher　recognition　accuracy.　In　addition,　we　analyze　the　impact　of　different　network　structures　on　the　recognition　performance.　The　results　show　that　the　proposed　method　can　not　only　improve　the　recognition　accuracy,　but　also　have　simple　structure　and　better　real-time　performance.　©　2022　IEEE.