darkne可视化网络层（卷继层）源码解读（寻码溯源）

1. 可视化语句

darknet.exe visualize   cfg/yolo-obj.cfg .\backup\yolo-obj_last.weights

2. 调用函数一：可视化cfg和weight

C:darknet-master\src\darknet.c

void visualize(char *cfgfile, char *weightfile)
{
    network net = parse_network_cfg(cfgfile);//解析配置文件
    if(weightfile){
        load_weights(&net, weightfile);//加载权重模型至net网络中
    }
    visualize_network(net);
#ifdef OPENCV
    wait_until_press_key_cv();
#endif
}

3.调用函数二：可视化网络
C:\darknet-master\src\network.c

void visualize_network(network net)
{
    image *prev = 0;
    int i;
    char buff[256];
    for(i = 0; i < net.n; ++i){
        sprintf(buff, "Layer %d", i);
        layer l = net.layers[i];
        if(l.type == CONVOLUTIONAL){
            prev = visualize_convolutional_layer(l, buff, prev);
        }
    }
}

4.调用函数三：可视化卷继层

C:\darknet-master\src\convolutional_layer.c

image *visualize_convolutional_layer(convolutional_layer l, char *window, image *prev_weights)
{
    image *single_weights = get_weights(l);
    show_images(single_weights, l.n, window);

    image delta = get_convolutional_image(l);
    image dc = collapse_image_layers(delta, 1);
    char buff[256];
    sprintf(buff, "%s: Output", window);
    //show_image(dc, buff);
    //save_image(dc, buff);
    free_image(dc);
    return single_weights;
}

5.调用函数四：获得权重

C:\darknet-master\src\convolutional_layer.c

image *get_weights(convolutional_layer l)
{
    image *weights = (image *)calloc(l.n, sizeof(image));
    int i;
    for (i = 0; i < l.n; ++i) {
        weights[i] = copy_image(get_convolutional_weight(l, i));
        normalize_image(weights[i]);
        /*
        char buff[256];
        sprintf(buff, "filter%d", i);
        save_image(weights[i], buff);
        */
    }
    //error("hey");
    return weights;
}

6.调用函数五：归一化
C:\darknet-master\src\image.c

void normalize_image(image p)
{
    int i;
    float min = 9999999;
    float max = -999999;

    for(i = 0; i < p.h*p.w*p.c; ++i){
        float v = p.data[i];
        if(v < min) min = v;
        if(v > max) max = v;
    }
    if(max - min < .000000001){
        min = 0;
        max = 1;
    }
    for(i = 0; i < p.c*p.w*p.h; ++i){
        p.data[i] = (p.data[i] - min)/(max-min);
    }
}