darknet/src/region_layer.c

#include "region_layer.h"
#include "activations.h"
#include "blas.h"
#include "box.h"
#include "cuda.h"
#include "utils.h"
#include <stdio.h>
#include <assert.h>
#include <string.h>
#include <stdlib.h>

layer make_region_layer(int batch, int w, int h, int n, int classes, int coords)
{
    layer l = {0};
    l.type = REGION;

    l.n = n;
    l.batch = batch;
    l.h = h;
    l.w = w;
    l.c = n*(classes + coords + 1);
    l.out_w = l.w;
    l.out_h = l.h;
    l.out_c = l.c;
    l.classes = classes;
    l.coords = coords;
    l.cost = calloc(1, sizeof(float));
    l.biases = calloc(n*2, sizeof(float));
    l.bias_updates = calloc(n*2, sizeof(float));
    l.outputs = h*w*n*(classes + coords + 1);
    l.inputs = l.outputs;
    l.truths = 30*(5);
    l.delta = calloc(batch*l.outputs, sizeof(float));
    l.output = calloc(batch*l.outputs, sizeof(float));
    int i;
    for(i = 0; i < n*2; ++i){
        l.biases[i] = .5;
    }

    l.forward = forward_region_layer;
    l.backward = backward_region_layer;
#ifdef GPU
    l.forward_gpu = forward_region_layer_gpu;
    l.backward_gpu = backward_region_layer_gpu;
    l.output_gpu = cuda_make_array(l.output, batch*l.outputs);
    l.delta_gpu = cuda_make_array(l.delta, batch*l.outputs);
#endif

    fprintf(stderr, "detection\n");
    srand(0);

    return l;
}

void resize_region_layer(layer *l, int w, int h)
{
    l->w = w;
    l->h = h;

    l->outputs = h*w*l->n*(l->classes + l->coords + 1);
    l->inputs = l->outputs;

    l->output = realloc(l->output, l->batch*l->outputs*sizeof(float));
    l->delta = realloc(l->delta, l->batch*l->outputs*sizeof(float));

#ifdef GPU
    cuda_free(l->delta_gpu);
    cuda_free(l->output_gpu);

    l->delta_gpu =     cuda_make_array(l->delta, l->batch*l->outputs);
    l->output_gpu =    cuda_make_array(l->output, l->batch*l->outputs);
#endif
}

box get_region_box(float *x, float *biases, int n, int index, int i, int j, int w, int h, int stride)
{
    box b;
    b.x = (i + x[index + 0*stride]) / w;
    b.y = (j + x[index + 1*stride]) / h;
    b.w = exp(x[index + 2*stride]) * biases[2*n]   / w;
    b.h = exp(x[index + 3*stride]) * biases[2*n+1] / h;
    return b;
}

float delta_region_box(box truth, float *x, float *biases, int n, int index, int i, int j, int w, int h, float *delta, float scale, int stride)
{
    box pred = get_region_box(x, biases, n, index, i, j, w, h, stride);
    float iou = box_iou(pred, truth);

    float tx = (truth.x*w - i);
    float ty = (truth.y*h - j);
    float tw = log(truth.w*w / biases[2*n]);
    float th = log(truth.h*h / biases[2*n + 1]);

    delta[index + 0*stride] = scale * (tx - x[index + 0*stride]);
    delta[index + 1*stride] = scale * (ty - x[index + 1*stride]);
    delta[index + 2*stride] = scale * (tw - x[index + 2*stride]);
    delta[index + 3*stride] = scale * (th - x[index + 3*stride]);
    return iou;
}

void delta_region_class(float *output, float *delta, int index, int class, int classes, tree *hier, float scale, int stride, float *avg_cat)
{
    int i, n;
    if(hier){
        float pred = 1;
        while(class >= 0){
            pred *= output[index + stride*class];
            int g = hier->group[class];
            int offset = hier->group_offset[g];
            for(i = 0; i < hier->group_size[g]; ++i){
                delta[index + stride*(offset + i)] = scale * (0 - output[index + stride*(offset + i)]);
            }
            delta[index + stride*class] = scale * (1 - output[index + stride*class]);

            class = hier->parent[class];
        }
        *avg_cat += pred;
    } else {
        for(n = 0; n < classes; ++n){
            delta[index + stride*n] = scale * (((n == class)?1 : 0) - output[index + stride*n]);
            if(n == class) *avg_cat += output[index + stride*n];
        }
    }
}

float logit(float x)
{
    return log(x/(1.-x));
}

float tisnan(float x)
{
    return (x != x);
}

int entry_index(layer l, int batch, int location, int entry)
{
    int n =   location / (l.w*l.h);
    int loc = location % (l.w*l.h);
    return batch*l.outputs + n*l.w*l.h*(l.coords+l.classes+1) + entry*l.w*l.h + loc;
}

void softmax_tree(float *input, int batch, int inputs, float temp, tree *hierarchy, float *output);
void forward_region_layer(const layer l, network net)
{
    int i,j,b,t,n;
    memcpy(l.output, net.input, l.outputs*l.batch*sizeof(float));

#ifndef GPU
    for (b = 0; b < l.batch; ++b){
        for(n = 0; n < l.n; ++n){
            int index = entry_index(l, b, n*l.w*l.h, 0);
            activate_array(l.output + index, 2*l.w*l.h, LOGISTIC);
            index = entry_index(l, b, n*l.w*l.h, 4);
            activate_array(l.output + index,   l.w*l.h, LOGISTIC);
        }
    }
    if (l.softmax_tree){
        int i;
        int count = 5;
        for (i = 0; i < l.softmax_tree->groups; ++i) {
            int group_size = l.softmax_tree->group_size[i];
            softmax_cpu(net.input + count, group_size, l.batch, l.inputs, l.n*l.w*l.h, 1, l.n*l.w*l.h, l.temperature, l.output + count);
            count += group_size;
        }
    } else if (l.softmax){
        int index = entry_index(l, 0, 0, 5);
        softmax_cpu(net.input + index, l.classes, l.batch*l.n, l.inputs/l.n, l.w*l.h, 1, l.w*l.h, 1, l.output + index);
    }
#endif

    memset(l.delta, 0, l.outputs * l.batch * sizeof(float));
    if(!net.train) return;
    float avg_iou = 0;
    float recall = 0;
    float avg_cat = 0;
    float avg_obj = 0;
    float avg_anyobj = 0;
    int count = 0;
    int class_count = 0;
    *(l.cost) = 0;
    for (b = 0; b < l.batch; ++b) {
        if(l.softmax_tree){
            int onlyclass = 0;
            for(t = 0; t < 30; ++t){
                box truth = float_to_box(net.truth + t*5 + b*l.truths, 1);
                if(!truth.x) break;
                int class = net.truth[t*5 + b*l.truths + 4];
                float maxp = 0;
                int maxi = 0;
                if(truth.x > 100000 && truth.y > 100000){
                    for(n = 0; n < l.n*l.w*l.h; ++n){
                        int class_index = entry_index(l, b, n, 5);
                        int obj_index = entry_index(l, b, n, 4);
                        float scale =  l.output[obj_index];
                        l.delta[obj_index] = l.noobject_scale * (0 - l.output[obj_index]);
                        float p = scale*get_hierarchy_probability(l.output + class_index, l.softmax_tree, class, l.w*l.h);
                        if(p > maxp){
                            maxp = p;
                            maxi = n;
                        }
                    }
                    int class_index = entry_index(l, b, maxi, 5);
                    int obj_index = entry_index(l, b, maxi, 4);
                    delta_region_class(l.output, l.delta, class_index, class, l.classes, l.softmax_tree, l.class_scale, l.w*l.h, &avg_cat);
                    if(l.output[obj_index] < .3) l.delta[obj_index] = l.object_scale * (.3 - l.output[obj_index]);
                    else  l.delta[obj_index] = 0;
                    ++class_count;
                    onlyclass = 1;
                    break;
                }
            }
            if(onlyclass) continue;
        }
        for (j = 0; j < l.h; ++j) {
            for (i = 0; i < l.w; ++i) {
                for (n = 0; n < l.n; ++n) {
                    int box_index = entry_index(l, b, n*l.w*l.h + j*l.w + i, 0);
                    box pred = get_region_box(l.output, l.biases, n, box_index, i, j, l.w, l.h, l.w*l.h);
                    float best_iou = 0;
                    for(t = 0; t < 30; ++t){
                        box truth = float_to_box(net.truth + t*5 + b*l.truths, 1);
                        if(!truth.x) break;
                        float iou = box_iou(pred, truth);
                        if (iou > best_iou) {
                            best_iou = iou;
                        }
                    }
                    int obj_index = entry_index(l, b, n*l.w*l.h + j*l.w + i, 4);
                    avg_anyobj += l.output[obj_index];
                    l.delta[obj_index] = l.noobject_scale * (0 - l.output[obj_index]);
                    if (best_iou > l.thresh) {
                        l.delta[obj_index] = 0;
                    }

                    if(*(net.seen) < 12800){
                        box truth = {0};
                        truth.x = (i + .5)/l.w;
                        truth.y = (j + .5)/l.h;
                        truth.w = l.biases[2*n]/l.w;
                        truth.h = l.biases[2*n+1]/l.h;
                        delta_region_box(truth, l.output, l.biases, n, box_index, i, j, l.w, l.h, l.delta, .01, l.w*l.h);
                    }
                }
            }
        }
        for(t = 0; t < 30; ++t){
            box truth = float_to_box(net.truth + t*5 + b*l.truths, 1);

            if(!truth.x) break;
            float best_iou = 0;
            int best_n = 0;
            i = (truth.x * l.w);
            j = (truth.y * l.h);
            //printf("%d %f %d %f\n", i, truth.x*l.w, j, truth.y*l.h);
            box truth_shift = truth;
            truth_shift.x = 0;
            truth_shift.y = 0;
            //printf("index %d %d\n",i, j);
            for(n = 0; n < l.n; ++n){
                int box_index = entry_index(l, b, n*l.w*l.h + j*l.w + i, 0);
                box pred = get_region_box(l.output, l.biases, n, box_index, i, j, l.w, l.h, l.w*l.h);
                if(l.bias_match){
                    pred.w = l.biases[2*n]/l.w;
                    pred.h = l.biases[2*n+1]/l.h;
                }
                //printf("pred: (%f, %f) %f x %f\n", pred.x, pred.y, pred.w, pred.h);
                pred.x = 0;
                pred.y = 0;
                float iou = box_iou(pred, truth_shift);
                if (iou > best_iou){
                    best_iou = iou;
                    best_n = n;
                }
            }
            //printf("%d %f (%f, %f) %f x %f\n", best_n, best_iou, truth.x, truth.y, truth.w, truth.h);

            int box_index = entry_index(l, b, best_n*l.w*l.h + j*l.w + i, 0);
            float iou = delta_region_box(truth, l.output, l.biases, best_n, box_index, i, j, l.w, l.h, l.delta, l.coord_scale *  (2 - truth.w*truth.h), l.w*l.h);
            if(iou > .5) recall += 1;
            avg_iou += iou;

            //l.delta[best_index + 4] = iou - l.output[best_index + 4];
            int obj_index = entry_index(l, b, best_n*l.w*l.h + j*l.w + i, 4);
            avg_obj += l.output[obj_index];
            l.delta[obj_index] = l.object_scale * (1 - l.output[obj_index]);
            if (l.rescore) {
                l.delta[obj_index] = l.object_scale * (iou - l.output[obj_index]);
            }

            int class = net.truth[t*5 + b*l.truths + 4];
            if (l.map) class = l.map[class];
            int class_index = entry_index(l, b, best_n*l.w*l.h + j*l.w + i, 5);
            delta_region_class(l.output, l.delta, class_index, class, l.classes, l.softmax_tree, l.class_scale, l.w*l.h, &avg_cat);
            ++count;
            ++class_count;
        }
    }
    //printf("\n");
    *(l.cost) = pow(mag_array(l.delta, l.outputs * l.batch), 2);
    printf("Region Avg IOU: %f, Class: %f, Obj: %f, No Obj: %f, Avg Recall: %f,  count: %d\n", avg_iou/count, avg_cat/class_count, avg_obj/count, avg_anyobj/(l.w*l.h*l.n*l.batch), recall/count, count);
}

void backward_region_layer(const layer l, network net)
{
    /*
       int b;
       int size = l.coords + l.classes + 1;
       for (b = 0; b < l.batch*l.n; ++b){
       int index = (b*size + 4)*l.w*l.h;
       gradient_array(l.output + index, l.w*l.h, LOGISTIC, l.delta + index);
       }
       axpy_cpu(l.batch*l.inputs, 1, l.delta, 1, net.delta, 1);
     */
}

void get_region_boxes(layer l, int w, int h, float thresh, float **probs, box *boxes, int only_objectness, int *map, float tree_thresh, int nomult)
{
    int i,j,n,z;
    float *predictions = l.output;
    if (l.batch == 2) {
        float *flip = l.output + l.outputs;
        for (j = 0; j < l.h; ++j) {
            for (i = 0; i < l.w/2; ++i) {
                for (n = 0; n < l.n; ++n) {
                    for(z = 0; z < l.classes + 5; ++z){
                        int i1 = z*l.w*l.h*l.n + n*l.w*l.h + j*l.w + i;
                        int i2 = z*l.w*l.h*l.n + n*l.w*l.h + j*l.w + (l.w - i - 1);
                        float swap = flip[i1];
                        flip[i1] = flip[i2];
                        flip[i2] = swap;
                        if(z == 0){
                            flip[i1] = -flip[i1];
                            flip[i2] = -flip[i2];
                        }
                    }
                }
            }
        }
        for(i = 0; i < l.outputs; ++i){
            l.output[i] = (l.output[i] + flip[i])/2.;
        }
    }
    for (i = 0; i < l.w*l.h; ++i){
        int row = i / l.w;
        int col = i % l.w;
        for(n = 0; n < l.n; ++n){
            int index = n*l.w*l.h + i;
            for(j = 0; j < l.classes; ++j){
                probs[index][j] = 0;
            }
            int obj_index = entry_index(l, 0, n*l.w*l.h + i, 4);
            int box_index = entry_index(l, 0, n*l.w*l.h + i, 0);
            float scale = predictions[obj_index];
            boxes[index] = get_region_box(predictions, l.biases, n, box_index, col, row, l.w, l.h, l.w*l.h);
            if(1){
                int max = w > h ? w : h;
                boxes[index].x =  (boxes[index].x - (max - w)/2./max) / ((float)w/max); 
                boxes[index].y =  (boxes[index].y - (max - h)/2./max) / ((float)h/max); 
                boxes[index].w *= (float)max/w;
                boxes[index].h *= (float)max/h;
            }
            if(!nomult){
                boxes[index].x *= w;
                boxes[index].y *= h;
                boxes[index].w *= w;
                boxes[index].h *= h;
            }

            int class_index = entry_index(l, 0, n*l.w*l.h + i, 5);
            if(l.softmax_tree){

                hierarchy_predictions(predictions + class_index, l.classes, l.softmax_tree, 0, l.w*l.h);
                if(map){
                    for(j = 0; j < 200; ++j){
                        int class_index = entry_index(l, 0, n*l.w*l.h + i, 5 + map[j]);
                        float prob = scale*predictions[class_index];
                        probs[index][j] = (prob > thresh) ? prob : 0;
                    }
                } else {
                    int j =  hierarchy_top_prediction(predictions + class_index, l.softmax_tree, tree_thresh, l.w*l.h);
                    probs[index][j] = (scale > thresh) ? scale : 0;
                    probs[index][l.classes] = scale;
                }
            } else {
                float max = 0;
                for(j = 0; j < l.classes; ++j){
                    int class_index = entry_index(l, 0, n*l.w*l.h + i, 5 + j);
                    float prob = scale*predictions[class_index];
                    probs[index][j] = (prob > thresh) ? prob : 0;
                    if(prob > max) max = prob;
                    // TODO REMOVE
                    // if (j != 15 && j != 16) probs[index][j] = 0; 
                    /*
                     if (j != 0) probs[index][j] = 0; 
                     int blacklist[] = {121, 497, 482, 504, 122, 518,481, 418, 542, 491, 914, 478, 120, 510,500};
                     int bb;
                     for (bb = 0; bb < sizeof(blacklist)/sizeof(int); ++bb){
                        if(index == blacklist[bb]) probs[index][j] = 0;
                     }
                     */
                }
                probs[index][l.classes] = max;
            }
            if(only_objectness){
                probs[index][0] = scale;
            }
        }
    }
}

#ifdef GPU

void forward_region_layer_gpu(const layer l, network net)
{
    copy_ongpu(l.batch*l.inputs, net.input_gpu, 1, l.output_gpu, 1);
    int b, n;
    for (b = 0; b < l.batch; ++b){
        for(n = 0; n < l.n; ++n){
            int index = entry_index(l, b, n*l.w*l.h, 0);
            activate_array_ongpu(l.output_gpu + index, 2*l.w*l.h, LOGISTIC);
            index = entry_index(l, b, n*l.w*l.h, 4);
            activate_array_ongpu(l.output_gpu + index,   l.w*l.h, LOGISTIC);
        }
    }
    if (l.softmax_tree){
        int i;
        int count = 5;
        for (i = 0; i < l.softmax_tree->groups; ++i) {
            int group_size = l.softmax_tree->group_size[i];
            int index = entry_index(l, 0, 0, count);
            softmax_gpu(net.input_gpu + index, group_size, l.batch*l.n, l.inputs/l.n, l.w*l.h, 1, l.w*l.h, 1, l.output_gpu + index);
            count += group_size;
        }
    } else if (l.softmax) {
        int index = entry_index(l, 0, 0, 5);
        //printf("%d\n", index);
        softmax_gpu(net.input_gpu + index, l.classes, l.batch*l.n, l.inputs/l.n, l.w*l.h, 1, l.w*l.h, 1, l.output_gpu + index);
    }
    if(!net.train || l.onlyforward){
        cuda_pull_array(l.output_gpu, l.output, l.batch*l.outputs);
        return;
    }

    float *truth_cpu = 0;
    if(net.truth_gpu){
        int num_truth = l.batch*l.truths;
        truth_cpu = calloc(num_truth, sizeof(float));
        cuda_pull_array(net.truth_gpu, truth_cpu, num_truth);
    }
    cuda_pull_array(l.output_gpu, net.input, l.batch*l.inputs);
    forward_region_layer(l, net);
    //cuda_push_array(l.output_gpu, l.output, l.batch*l.outputs);
    if(!net.train) return;
    cuda_push_array(l.delta_gpu, l.delta, l.batch*l.outputs);
}

void backward_region_layer_gpu(const layer l, network net)
{
    int b, n;
    for (b = 0; b < l.batch; ++b){
        for(n = 0; n < l.n; ++n){
            int index = entry_index(l, b, n*l.w*l.h, 0);
            gradient_array_ongpu(l.output_gpu + index, 2*l.w*l.h, LOGISTIC, l.delta_gpu + index);
            index = entry_index(l, b, n*l.w*l.h, 4);
            gradient_array_ongpu(l.output_gpu + index,   l.w*l.h, LOGISTIC, l.delta_gpu + index);
        }
    }
    axpy_ongpu(l.batch*l.inputs, 1, l.delta_gpu, 1, net.delta_gpu, 1);
}
#endif

void zero_objectness(layer l)
{
    int i, n;
    for (i = 0; i < l.w*l.h; ++i){
        for(n = 0; n < l.n; ++n){
            int obj_index = entry_index(l, 0, n*l.w*l.h + i, 4);
            l.output[obj_index] = 0;
        }
    }
}