YOLO回归函数代码对应讲解

最新推荐文章于 2025-05-29 15:32:29 发布

原创最新推荐文章于 2025-05-29 15:32:29 发布 · 1.6k 阅读

3 ·

CC 4.0 BY-SA版权

文章标签：

#YOLO代码理论简略讲解

DeepLearning 专栏收录该内容

18 篇文章

订阅专栏

本文详细解析了YOLO目标检测算法的工作原理，包括其回归函数的具体实现方式，并探讨了网络训练过程中涉及的关键变量与参数调整策略。此外，还讨论了YOLO在处理重叠目标时的局限性。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

关于YOLO的理论知识个人转载的一篇博客有具体讲解：点击打开链接，本文主要对YOLO的回归函数与代码对应起来讲解，便于更加深入的理解YOLO的实现过程。

if(state.train){
        float avg_iou = 0;
        float avg_cat = 0;
        float avg_allcat = 0;
        float avg_obj = 0;
        float avg_anyobj = 0;
        int count = 0;
        *(l.cost) = 0;
        int size = l.inputs * l.batch;
        memset(l.delta, 0, size * sizeof(float));
        for (b = 0; b < l.batch; ++b){
            int index = b*l.inputs;
            for (i = 0; i < locations; ++i) {
                int truth_index = (b*locations + i)*(1+l.coords+l.classes);
                int is_obj = state.truth[truth_index];//表示这个cell是否是object，数据存放在state里面
                for (j = 0; j < l.n; ++j) {
                    int p_index = index + locations*l.classes + i*l.n + j;
                    l.delta[p_index] = l.noobject_scale*(0 - l.output[p_index]);//没有目标物体项
                    *(l.cost) += l.noobject_scale*pow(l.output[p_index], 2);
                    avg_anyobj += l.output[p_index];
                }

                int best_index = -1;
                float best_iou = 0;
                float best_rmse = 20;

                if (!is_obj){
                    continue;
                }

		//类别项
                int class_index = index + i*l.classes;
                for(j = 0; j < l.classes; ++j) {
                    l.delta[class_index+j] = l.class_scale * (state.truth[truth_index+1+j] - l.output[class_index+j]);
                    *(l.cost) += l.class_scale * pow(state.truth[truth_index+1+j] - l.output[class_index+j], 2);
                    if(state.truth[truth_index + 1 + j]) avg_cat += l.output[class_index+j];
                    avg_allcat += l.output[class_index+j];
                }

                box truth = float_to_box(state.truth + truth_index + 1 + l.classes);
                truth.x /= l.side;
                truth.y /= l.side;

                for(j = 0; j < l.n; ++j){
                    int box_index = index + locations*(l.classes + l.n) + (i*l.n + j) * l.coords;
                    box out = float_to_box(l.output + box_index);
                    out.x /= l.side;
                    out.y /= l.side;

                    if (l.sqrt){
                        out.w = out.w*out.w;
                        out.h = out.h*out.h;
                    }

                    float iou  = box_iou(out, truth);
                    //iou = 0;
                    float rmse = box_rmse(out, truth);
                    if(best_iou > 0 || iou > 0){
                        if(iou > best_iou){
                            best_iou = iou;
                            best_index = j;
                        }
                    }else{
                        if(rmse < best_rmse){
                            best_rmse = rmse;
                            best_index = j;
                        }
                    }
                }

                if(l.forced){
                    if(truth.w*truth.h < .1){
                        best_index = 1;
                    }else{
                        best_index = 0;
                    }
                }
                if(l.random && *(state.net.seen) < 64000){
                    best_index = rand()%l.n;
                }

                int box_index = index + locations*(l.classes + l.n) + (i*l.n + best_index) * l.coords;
                int tbox_index = truth_index + 1 + l.classes;

                box out = float_to_box(l.output + box_index);
                out.x /= l.side;
                out.y /= l.side;
                if (l.sqrt) {
                    out.w = out.w*out.w;
                    out.h = out.h*out.h;
                }
                float iou  = box_iou(out, truth);

                //printf("%d,", best_index);
		//对应于论文里面的第三项
                int p_index = index + locations*l.classes + i*l.n + best_index;//这里揭开前面的问题，这里相当于已经判定是objcet，就把之前算在里面noobjcet的减掉
                *(l.cost) -= l.noobject_scale * pow(l.output[p_index], 2);
                *(l.cost) += l.object_scale * pow(1-l.output[p_index], 2);
                avg_obj += l.output[p_index];
                l.delta[p_index] = l.object_scale * (1.-l.output[p_index]);//这里就覆盖到之前被noobject赋值的数

                if(l.rescore){
                    l.delta[p_index] = l.object_scale * (iou - l.output[p_index]);
                }

		//下面的四个对应于论文里面损失的第一项和第二项, 即坐标的回归。
                l.delta[box_index+0] = l.coord_scale*(state.truth[tbox_index + 0] - l.output[box_index + 0]);//deltax
                l.delta[box_index+1] = l.coord_scale*(state.truth[tbox_index + 1] - l.output[box_index + 1]);//deltay
                l.delta[box_index+2] = l.coord_scale*(state.truth[tbox_index + 2] - l.output[box_index + 2]);//deltaw
                l.delta[box_index+3] = l.coord_scale*(state.truth[tbox_index + 3] - l.output[box_index + 3]);//deltah
                if(l.sqrt){
                    l.delta[box_index+2] = l.coord_scale*(sqrt(state.truth[tbox_index + 2]) - l.output[box_index + 2]);//sqrt-deltaw;
                    l.delta[box_index+3] = l.coord_scale*(sqrt(state.truth[tbox_index + 3]) - l.output[box_index + 3]);//sqrt-deltah;
                }

                *(l.cost) += pow(1-iou, 2);
                avg_iou += iou;
                ++count;
            }
        }

从detect_layer.c中网络训练的代码发现，目标物体的GT维度为25，分别为类别维度(20)，置信度(1)，坐标(4)；yolo实质上每格小格只预测一个box，因为仅选择预测到的两个box中IOU较大的作为最终的box回归。通过这种回归模式，显然对于目标物体存在重叠度较大，或者目标物体较为密集的情形下，YOLO的这种预测模式会存在较大的漏检，效果有待进一步改进。网络输出数据格式和标签的格式如下图所示：

若有不当之处，请指教，谢谢！