Python基于HOG+SVM/RF/DT等模型实现目标人行检测功能

def img2Feature(dataDir="data/",save_path="feature.json"):
    """
    特征提取计算
    """
    feature=[]
    for one_label in os.listdir(dataDir):
        print("one_label: ", one_label)
        oneDir=dataDir+one_label+'/'
        for one_pic in os.listdir(oneDir):
            one_path=oneDir+one_pic
            print("one_path: ", one_path)
            #加载图像
            one_img = imread(one_path, as_gray=True)
            one_vec = hog(one_img, orientations=orientations, pixels_per_cell=pixels_per_cell, cells_per_block=cells_per_block, 
                        visualize=visualize, block_norm=normalize)
            one_vec=one_vec.tolist()
            one_vec.append(one_label)
            feature.append(one_vec)
    print("feature_length: ", len(feature))
    with open(save_path,"w") as f:
        f.write(json.dumps(feature))

resDir = "results/"
if not os.path.exists(resDir):
    os.makedirs(resDir)
data = "feature.json"
dict1 = DTModel(data=data, rationum=0.25, model_path=resDir + "DT.model")
dict2 = RFModel(data=data, rationum=0.25, model_path=resDir + "RF.model")
dict3 = SVMModel(data=data, rationum=0.25, model_path=resDir + "SVM.model")
res_dict = {}
res_dict["DT"], res_dict["RF"], res_dict["SVM"] = dict1, dict2, dict3
with open(resDir + "res_dict.json", "w") as f:
    f.write(json.dumps(res_dict))
comparePloter(dict1, dict2, dict3, save_path=resDir + "comparePloter.jpg")

{
	"DT": {
		"precision": 0.7573482282561567,
		"recall": 0.7597846737437716,
		"F1": 0.7584933696379963,
		"accuracy": 0.7584933696379963
	},
	"RF": {
		"precision": 0.9156160607479066,
		"recall": 0.8801773928046967,
		"F1": 0.893107332148193,
		"accuracy": 0.893107332148193
	},
	"SVM": {
		"precision": 0.9281402443868877,
		"recall": 0.9272928963585789,
		"F1": 0.9277128372009962,
		"accuracy": 0.9277128372009962
	}
}