The　infrared　heat　map　reflects　the　overall　temperature　distribution　of　the　human　body,　which　coincides　with　the　theory　of　traditional　Chinese　medicine,　as　an　important　indicator　of　human　sub-health　status　identification　in　the　field　of　traditional　Chinese　medicine,　the　physical　fitness　of　traditional　Chinese　medicine　has　attracted　more　and　more　attention　from　the　medical　field　and　the　general　public.　Because　the　infrared　heat　map　of　the　human　body　has　rich　color　distribution　characteristics,　this　paper　uses　color　feature　extraction　and　texture　feature　extraction　algorithms　combined　with　SVM　classifiers　and　convolutional　neural　networks　to　perform　classification　experiments.　The　results　show　that　the　accuracy　of　deep　learning　algorithms　is　higher　than　that　of　traditional　machine　learning　algorithms,　Deep　learning　can　be　combined　with　local　subtle　features　for　further　research.　©　Published　under　licence　by　IOP　Publishing　Ltd.