In　recent　years,　with　the　gradual　deepening　of　deep　learning　technology,　medical　image　classification　technology　has　had　a　rapid　development,　and　a　number　of　classical　models　and　algorithms　have　emerged　one　after　another.　In　this　paper,　from　the　perspective　of　the　technical　development　of　image　classification,　it　is　divided　into　two　categories:　traditional　technology　and　deep　learning-based　technology.　Focusing　on　the　typical　classification　network　architecture　based　on　deep　learning,　a　comprehensive　analysis　is　made,　and　the　processing　approaches　proposed　for　medical　images　in　recent　years　are　sorted　out.