diff --git a/cfg/writing.cfg b/cfg/writing.cfg
new file mode 100644
index 0000000000000000000000000000000000000000..2c84b197d92cdb147174c5bcf8d5194af7abf721
--- /dev/null
+++ b/cfg/writing.cfg
@@ -0,0 +1,49 @@
+[net]
+batch=64
+subdivisions=1
+height=256
+width=256
+channels=3
+learning_rate=0.00001
+momentum=0.9
+decay=0.0005
+seen=0
+
+[crop]
+crop_height=256
+crop_width=256
+flip=0
+angle=0
+saturation=1
+exposure=1
+
+[convolutional]
+filters=32
+size=3
+stride=1
+pad=1
+activation=ramp
+
+[convolutional]
+filters=32
+size=3
+stride=1
+pad=1
+activation=ramp
+
+[convolutional]
+filters=32
+size=3
+stride=1
+pad=1
+activation=ramp
+
+[convolutional]
+filters=1
+size=5
+stride=1
+pad=1
+activation=logistic
+
+[cost]
+
diff --git a/src/data.c b/src/data.c
index 17772c1c5625cc0bbe299e902f7ca2863f0ae173..7ef154663155786d232264ad4d09082755107505 100644
--- a/src/data.c
+++ b/src/data.c
@@ -54,7 +54,12 @@ matrix load_image_paths_gray(char **paths, int n, int w, int h)
X.cols = 0;
for(i = 0; i < n; ++i){
- image im = load_image(paths[i], w, h, 1);
+ image im = load_image(paths[i], w, h, 3);
+
+ image gray = grayscale_image(im);
+ free_image(im);
+ im = gray;
+
X.vals[i] = im.data;
X.cols = im.h*im.w*im.c;
}
@@ -571,14 +576,14 @@ pthread_t load_data_in_thread(load_args args)
return thread;
}
-data load_data_writing(char **paths, int n, int m, int w, int h)
+data load_data_writing(char **paths, int n, int m, int w, int h, int downsample)
{
if(m) paths = get_random_paths(paths, n, m);
char **replace_paths = find_replace_paths(paths, n, ".png", "-label.png");
data d;
d.shallow = 0;
d.X = load_image_paths(paths, n, w, h);
- d.y = load_image_paths_gray(replace_paths, n, w/8, h/8);
+ d.y = load_image_paths_gray(replace_paths, n, w/downsample, h/downsample);
if(m) free(paths);
int i;
for(i = 0; i < n; ++i) free(replace_paths[i]);
diff --git a/src/data.h b/src/data.h
index 216ab0c4bd0379f5f74031d83c34bc44acbc7556..1343460cba2bc5b3fe7a81acec0389a161d0234a 100644
--- a/src/data.h
+++ b/src/data.h
@@ -68,7 +68,7 @@ box_label *read_boxes(char *filename, int *n);
data load_cifar10_data(char *filename);
data load_all_cifar10();
-data load_data_writing(char **paths, int n, int m, int w, int h);
+data load_data_writing(char **paths, int n, int m, int w, int h, int downsample);
list *get_paths(char *filename);
char **get_labels(char *filename);
diff --git a/src/image.c b/src/image.c
index b6d757788c6e06ac502cb68882df5b6786cd3ad8..fa48678eee7615d1964a74ae5a588a98453d34f3 100644
--- a/src/image.c
+++ b/src/image.c
@@ -241,650 +241,648 @@ void show_image_cv(image p, char *name)
}
cvShowImage(buff, disp);
cvReleaseImage(&disp);
- }
+}
#endif
- void show_image(image p, char *name)
- {
+void show_image(image p, char *name)
+{
#ifdef OPENCV
- show_image_cv(p, name);
+ show_image_cv(p, name);
#else
- fprintf(stderr, "Not compiled with OpenCV, saving to %s.png instead\n", name);
- save_image(p, name);
+ fprintf(stderr, "Not compiled with OpenCV, saving to %s.png instead\n", name);
+ save_image(p, name);
#endif
- }
+}
- void save_image(image im, char *name)
- {
- char buff[256];
- //sprintf(buff, "%s (%d)", name, windows);
- sprintf(buff, "%s.png", name);
- unsigned char *data = calloc(im.w*im.h*im.c, sizeof(char));
- int i,k;
- for(k = 0; k < im.c; ++k){
- for(i = 0; i < im.w*im.h; ++i){
- data[i*im.c+k] = (unsigned char) (255*im.data[i + k*im.w*im.h]);
- }
+void save_image(image im, char *name)
+{
+ char buff[256];
+ //sprintf(buff, "%s (%d)", name, windows);
+ sprintf(buff, "%s.png", name);
+ unsigned char *data = calloc(im.w*im.h*im.c, sizeof(char));
+ int i,k;
+ for(k = 0; k < im.c; ++k){
+ for(i = 0; i < im.w*im.h; ++i){
+ data[i*im.c+k] = (unsigned char) (255*im.data[i + k*im.w*im.h]);
}
- int success = stbi_write_png(buff, im.w, im.h, im.c, data, im.w*im.c);
- free(data);
- if(!success) fprintf(stderr, "Failed to write image %s\n", buff);
}
+ int success = stbi_write_png(buff, im.w, im.h, im.c, data, im.w*im.c);
+ free(data);
+ if(!success) fprintf(stderr, "Failed to write image %s\n", buff);
+}
#ifdef OPENCV
- void save_image_jpg(image p, char *name)
- {
- image copy = copy_image(p);
- rgbgr_image(copy);
- int x,y,k;
-
- char buff[256];
- sprintf(buff, "%s.jpg", name);
-
- IplImage *disp = cvCreateImage(cvSize(p.w,p.h), IPL_DEPTH_8U, p.c);
- int step = disp->widthStep;
- for(y = 0; y < p.h; ++y){
- for(x = 0; x < p.w; ++x){
- for(k= 0; k < p.c; ++k){
- disp->imageData[y*step + x*p.c + k] = (unsigned char)(get_pixel(copy,x,y,k)*255);
- }
+void save_image_jpg(image p, char *name)
+{
+ image copy = copy_image(p);
+ rgbgr_image(copy);
+ int x,y,k;
+
+ char buff[256];
+ sprintf(buff, "%s.jpg", name);
+
+ IplImage *disp = cvCreateImage(cvSize(p.w,p.h), IPL_DEPTH_8U, p.c);
+ int step = disp->widthStep;
+ for(y = 0; y < p.h; ++y){
+ for(x = 0; x < p.w; ++x){
+ for(k= 0; k < p.c; ++k){
+ disp->imageData[y*step + x*p.c + k] = (unsigned char)(get_pixel(copy,x,y,k)*255);
}
}
- cvSaveImage(buff, disp,0);
- cvReleaseImage(&disp);
- free_image(copy);
}
- #endif
+ cvSaveImage(buff, disp,0);
+ cvReleaseImage(&disp);
+ free_image(copy);
+}
+#endif
- void show_image_layers(image p, char *name)
- {
- int i;
- char buff[256];
- for(i = 0; i < p.c; ++i){
- sprintf(buff, "%s - Layer %d", name, i);
- image layer = get_image_layer(p, i);
- show_image(layer, buff);
- free_image(layer);
- }
+void show_image_layers(image p, char *name)
+{
+ int i;
+ char buff[256];
+ for(i = 0; i < p.c; ++i){
+ sprintf(buff, "%s - Layer %d", name, i);
+ image layer = get_image_layer(p, i);
+ show_image(layer, buff);
+ free_image(layer);
}
+}
- void show_image_collapsed(image p, char *name)
- {
- image c = collapse_image_layers(p, 1);
- show_image(c, name);
- free_image(c);
- }
+void show_image_collapsed(image p, char *name)
+{
+ image c = collapse_image_layers(p, 1);
+ show_image(c, name);
+ free_image(c);
+}
- image make_empty_image(int w, int h, int c)
- {
- image out;
- out.data = 0;
- out.h = h;
- out.w = w;
- out.c = c;
- return out;
- }
+image make_empty_image(int w, int h, int c)
+{
+ image out;
+ out.data = 0;
+ out.h = h;
+ out.w = w;
+ out.c = c;
+ return out;
+}
- image make_image(int w, int h, int c)
- {
- image out = make_empty_image(w,h,c);
- out.data = calloc(h*w*c, sizeof(float));
- return out;
- }
+image make_image(int w, int h, int c)
+{
+ image out = make_empty_image(w,h,c);
+ out.data = calloc(h*w*c, sizeof(float));
+ return out;
+}
- image float_to_image(int w, int h, int c, float *data)
- {
- image out = make_empty_image(w,h,c);
- out.data = data;
- return out;
- }
+image float_to_image(int w, int h, int c, float *data)
+{
+ image out = make_empty_image(w,h,c);
+ out.data = data;
+ return out;
+}
- image rotate_image(image im, float rad)
- {
- int x, y, c;
- float cx = im.w/2.;
- float cy = im.h/2.;
- image rot = make_image(im.w, im.h, im.c);
- for(c = 0; c < im.c; ++c){
- for(y = 0; y < im.h; ++y){
- for(x = 0; x < im.w; ++x){
- float rx = cos(rad)*(x-cx) - sin(rad)*(y-cy) + cx;
- float ry = sin(rad)*(x-cx) + cos(rad)*(y-cy) + cy;
- float val = bilinear_interpolate(im, rx, ry, c);
- set_pixel(rot, x, y, c, val);
- }
+image rotate_image(image im, float rad)
+{
+ int x, y, c;
+ float cx = im.w/2.;
+ float cy = im.h/2.;
+ image rot = make_image(im.w, im.h, im.c);
+ for(c = 0; c < im.c; ++c){
+ for(y = 0; y < im.h; ++y){
+ for(x = 0; x < im.w; ++x){
+ float rx = cos(rad)*(x-cx) - sin(rad)*(y-cy) + cx;
+ float ry = sin(rad)*(x-cx) + cos(rad)*(y-cy) + cy;
+ float val = bilinear_interpolate(im, rx, ry, c);
+ set_pixel(rot, x, y, c, val);
}
}
- return rot;
}
+ return rot;
+}
- void translate_image(image m, float s)
- {
- int i;
- for(i = 0; i < m.h*m.w*m.c; ++i) m.data[i] += s;
- }
+void translate_image(image m, float s)
+{
+ int i;
+ for(i = 0; i < m.h*m.w*m.c; ++i) m.data[i] += s;
+}
- void scale_image(image m, float s)
- {
- int i;
- for(i = 0; i < m.h*m.w*m.c; ++i) m.data[i] *= s;
- }
+void scale_image(image m, float s)
+{
+ int i;
+ for(i = 0; i < m.h*m.w*m.c; ++i) m.data[i] *= s;
+}
- image crop_image(image im, int dx, int dy, int w, int h)
- {
- image cropped = make_image(w, h, im.c);
- int i, j, k;
- for(k = 0; k < im.c; ++k){
- for(j = 0; j < h; ++j){
- for(i = 0; i < w; ++i){
- int r = j + dy;
- int c = i + dx;
- float val = 0;
- if (r >= 0 && r < im.h && c >= 0 && c < im.w) {
- val = get_pixel(im, c, r, k);
- }
- set_pixel(cropped, i, j, k, val);
+image crop_image(image im, int dx, int dy, int w, int h)
+{
+ image cropped = make_image(w, h, im.c);
+ int i, j, k;
+ for(k = 0; k < im.c; ++k){
+ for(j = 0; j < h; ++j){
+ for(i = 0; i < w; ++i){
+ int r = j + dy;
+ int c = i + dx;
+ float val = 0;
+ if (r >= 0 && r < im.h && c >= 0 && c < im.w) {
+ val = get_pixel(im, c, r, k);
}
+ set_pixel(cropped, i, j, k, val);
}
}
- return cropped;
}
+ return cropped;
+}
- float three_way_max(float a, float b, float c)
- {
- return (a > b) ? ( (a > c) ? a : c) : ( (b > c) ? b : c) ;
- }
+float three_way_max(float a, float b, float c)
+{
+ return (a > b) ? ( (a > c) ? a : c) : ( (b > c) ? b : c) ;
+}
- float three_way_min(float a, float b, float c)
- {
- return (a < b) ? ( (a < c) ? a : c) : ( (b < c) ? b : c) ;
- }
+float three_way_min(float a, float b, float c)
+{
+ return (a < b) ? ( (a < c) ? a : c) : ( (b < c) ? b : c) ;
+}
- // http://www.cs.rit.edu/~ncs/color/t_convert.html
- void rgb_to_hsv(image im)
- {
- assert(im.c == 3);
- int i, j;
- float r, g, b;
- float h, s, v;
- for(j = 0; j < im.h; ++j){
- for(i = 0; i < im.w; ++i){
- r = get_pixel(im, i , j, 0);
- g = get_pixel(im, i , j, 1);
- b = get_pixel(im, i , j, 2);
- float max = three_way_max(r,g,b);
- float min = three_way_min(r,g,b);
- float delta = max - min;
- v = max;
- if(max == 0){
- s = 0;
- h = -1;
- }else{
- s = delta/max;
- if(r == max){
- h = (g - b) / delta;
- } else if (g == max) {
- h = 2 + (b - r) / delta;
- } else {
- h = 4 + (r - g) / delta;
- }
- if (h < 0) h += 6;
+// http://www.cs.rit.edu/~ncs/color/t_convert.html
+void rgb_to_hsv(image im)
+{
+ assert(im.c == 3);
+ int i, j;
+ float r, g, b;
+ float h, s, v;
+ for(j = 0; j < im.h; ++j){
+ for(i = 0; i < im.w; ++i){
+ r = get_pixel(im, i , j, 0);
+ g = get_pixel(im, i , j, 1);
+ b = get_pixel(im, i , j, 2);
+ float max = three_way_max(r,g,b);
+ float min = three_way_min(r,g,b);
+ float delta = max - min;
+ v = max;
+ if(max == 0){
+ s = 0;
+ h = -1;
+ }else{
+ s = delta/max;
+ if(r == max){
+ h = (g - b) / delta;
+ } else if (g == max) {
+ h = 2 + (b - r) / delta;
+ } else {
+ h = 4 + (r - g) / delta;
}
- set_pixel(im, i, j, 0, h);
- set_pixel(im, i, j, 1, s);
- set_pixel(im, i, j, 2, v);
+ if (h < 0) h += 6;
}
+ set_pixel(im, i, j, 0, h);
+ set_pixel(im, i, j, 1, s);
+ set_pixel(im, i, j, 2, v);
}
}
+}
- void hsv_to_rgb(image im)
- {
- assert(im.c == 3);
- int i, j;
- float r, g, b;
- float h, s, v;
- float f, p, q, t;
- for(j = 0; j < im.h; ++j){
- for(i = 0; i < im.w; ++i){
- h = get_pixel(im, i , j, 0);
- s = get_pixel(im, i , j, 1);
- v = get_pixel(im, i , j, 2);
- if (s == 0) {
- r = g = b = v;
+void hsv_to_rgb(image im)
+{
+ assert(im.c == 3);
+ int i, j;
+ float r, g, b;
+ float h, s, v;
+ float f, p, q, t;
+ for(j = 0; j < im.h; ++j){
+ for(i = 0; i < im.w; ++i){
+ h = get_pixel(im, i , j, 0);
+ s = get_pixel(im, i , j, 1);
+ v = get_pixel(im, i , j, 2);
+ if (s == 0) {
+ r = g = b = v;
+ } else {
+ int index = floor(h);
+ f = h - index;
+ p = v*(1-s);
+ q = v*(1-s*f);
+ t = v*(1-s*(1-f));
+ if(index == 0){
+ r = v; g = t; b = p;
+ } else if(index == 1){
+ r = q; g = v; b = p;
+ } else if(index == 2){
+ r = p; g = v; b = t;
+ } else if(index == 3){
+ r = p; g = q; b = v;
+ } else if(index == 4){
+ r = t; g = p; b = v;
} else {
- int index = floor(h);
- f = h - index;
- p = v*(1-s);
- q = v*(1-s*f);
- t = v*(1-s*(1-f));
- if(index == 0){
- r = v; g = t; b = p;
- } else if(index == 1){
- r = q; g = v; b = p;
- } else if(index == 2){
- r = p; g = v; b = t;
- } else if(index == 3){
- r = p; g = q; b = v;
- } else if(index == 4){
- r = t; g = p; b = v;
- } else {
- r = v; g = p; b = q;
- }
+ r = v; g = p; b = q;
}
- set_pixel(im, i, j, 0, r);
- set_pixel(im, i, j, 1, g);
- set_pixel(im, i, j, 2, b);
}
+ set_pixel(im, i, j, 0, r);
+ set_pixel(im, i, j, 1, g);
+ set_pixel(im, i, j, 2, b);
}
}
+}
- image grayscale_image(image im)
- {
- assert(im.c == 3);
- int i, j, k;
- image gray = make_image(im.w, im.h, im.c);
- float scale[] = {0.587, 0.299, 0.114};
- for(k = 0; k < im.c; ++k){
- for(j = 0; j < im.h; ++j){
- for(i = 0; i < im.w; ++i){
- gray.data[i+im.w*j] += scale[k]*get_pixel(im, i, j, k);
- }
+image grayscale_image(image im)
+{
+ assert(im.c == 3);
+ int i, j, k;
+ image gray = make_image(im.w, im.h, 1);
+ float scale[] = {0.587, 0.299, 0.114};
+ for(k = 0; k < im.c; ++k){
+ for(j = 0; j < im.h; ++j){
+ for(i = 0; i < im.w; ++i){
+ gray.data[i+im.w*j] += scale[k]*get_pixel(im, i, j, k);
}
}
- memcpy(gray.data + im.w*im.h*1, gray.data, sizeof(float)*im.w*im.h);
- memcpy(gray.data + im.w*im.h*2, gray.data, sizeof(float)*im.w*im.h);
- return gray;
}
+ return gray;
+}
- image blend_image(image fore, image back, float alpha)
- {
- assert(fore.w == back.w && fore.h == back.h && fore.c == back.c);
- image blend = make_image(fore.w, fore.h, fore.c);
- int i, j, k;
- for(k = 0; k < fore.c; ++k){
- for(j = 0; j < fore.h; ++j){
- for(i = 0; i < fore.w; ++i){
- float val = alpha * get_pixel(fore, i, j, k) +
- (1 - alpha)* get_pixel(back, i, j, k);
- set_pixel(blend, i, j, k, val);
- }
+image blend_image(image fore, image back, float alpha)
+{
+ assert(fore.w == back.w && fore.h == back.h && fore.c == back.c);
+ image blend = make_image(fore.w, fore.h, fore.c);
+ int i, j, k;
+ for(k = 0; k < fore.c; ++k){
+ for(j = 0; j < fore.h; ++j){
+ for(i = 0; i < fore.w; ++i){
+ float val = alpha * get_pixel(fore, i, j, k) +
+ (1 - alpha)* get_pixel(back, i, j, k);
+ set_pixel(blend, i, j, k, val);
}
}
- return blend;
}
+ return blend;
+}
- void scale_image_channel(image im, int c, float v)
- {
- int i, j;
- for(j = 0; j < im.h; ++j){
- for(i = 0; i < im.w; ++i){
- float pix = get_pixel(im, i, j, c);
- pix = pix*v;
- set_pixel(im, i, j, c, pix);
- }
+void scale_image_channel(image im, int c, float v)
+{
+ int i, j;
+ for(j = 0; j < im.h; ++j){
+ for(i = 0; i < im.w; ++i){
+ float pix = get_pixel(im, i, j, c);
+ pix = pix*v;
+ set_pixel(im, i, j, c, pix);
}
}
+}
- void saturate_image(image im, float sat)
- {
- rgb_to_hsv(im);
- scale_image_channel(im, 1, sat);
- hsv_to_rgb(im);
- constrain_image(im);
- }
-
- void exposure_image(image im, float sat)
- {
- rgb_to_hsv(im);
- scale_image_channel(im, 2, sat);
- hsv_to_rgb(im);
- constrain_image(im);
- }
-
- void saturate_exposure_image(image im, float sat, float exposure)
- {
- rgb_to_hsv(im);
- scale_image_channel(im, 1, sat);
- scale_image_channel(im, 2, exposure);
- hsv_to_rgb(im);
- constrain_image(im);
- }
+void saturate_image(image im, float sat)
+{
+ rgb_to_hsv(im);
+ scale_image_channel(im, 1, sat);
+ hsv_to_rgb(im);
+ constrain_image(im);
+}
- /*
- image saturate_image(image im, float sat)
- {
- image gray = grayscale_image(im);
- image blend = blend_image(im, gray, sat);
- free_image(gray);
- constrain_image(blend);
- return blend;
- }
+void exposure_image(image im, float sat)
+{
+ rgb_to_hsv(im);
+ scale_image_channel(im, 2, sat);
+ hsv_to_rgb(im);
+ constrain_image(im);
+}
- image brightness_image(image im, float b)
- {
- image bright = make_image(im.w, im.h, im.c);
- return bright;
- }
- */
+void saturate_exposure_image(image im, float sat, float exposure)
+{
+ rgb_to_hsv(im);
+ scale_image_channel(im, 1, sat);
+ scale_image_channel(im, 2, exposure);
+ hsv_to_rgb(im);
+ constrain_image(im);
+}
- float bilinear_interpolate(image im, float x, float y, int c)
- {
- int ix = (int) floorf(x);
- int iy = (int) floorf(y);
+/*
+ image saturate_image(image im, float sat)
+ {
+ image gray = grayscale_image(im);
+ image blend = blend_image(im, gray, sat);
+ free_image(gray);
+ constrain_image(blend);
+ return blend;
+ }
+
+ image brightness_image(image im, float b)
+ {
+ image bright = make_image(im.w, im.h, im.c);
+ return bright;
+ }
+ */
+
+float bilinear_interpolate(image im, float x, float y, int c)
+{
+ int ix = (int) floorf(x);
+ int iy = (int) floorf(y);
- float dx = x - ix;
- float dy = y - iy;
+ float dx = x - ix;
+ float dy = y - iy;
- float val = (1-dy) * (1-dx) * get_pixel_extend(im, ix, iy, c) +
- dy * (1-dx) * get_pixel_extend(im, ix, iy+1, c) +
- (1-dy) * dx * get_pixel_extend(im, ix+1, iy, c) +
- dy * dx * get_pixel_extend(im, ix+1, iy+1, c);
- return val;
- }
+ float val = (1-dy) * (1-dx) * get_pixel_extend(im, ix, iy, c) +
+ dy * (1-dx) * get_pixel_extend(im, ix, iy+1, c) +
+ (1-dy) * dx * get_pixel_extend(im, ix+1, iy, c) +
+ dy * dx * get_pixel_extend(im, ix+1, iy+1, c);
+ return val;
+}
- image resize_image(image im, int w, int h)
- {
- image resized = make_image(w, h, im.c);
- image part = make_image(w, im.h, im.c);
- int r, c, k;
- float w_scale = (float)(im.w - 1) / (w - 1);
- float h_scale = (float)(im.h - 1) / (h - 1);
- for(k = 0; k < im.c; ++k){
- for(r = 0; r < im.h; ++r){
- for(c = 0; c < w; ++c){
- float val = 0;
- if(c == w-1){
- val = get_pixel(im, im.w-1, r, k);
- } else {
- float sx = c*w_scale;
- int ix = (int) sx;
- float dx = sx - ix;
- val = (1 - dx) * get_pixel(im, ix, r, k) + dx * get_pixel(im, ix+1, r, k);
- }
- set_pixel(part, c, r, k, val);
+image resize_image(image im, int w, int h)
+{
+ image resized = make_image(w, h, im.c);
+ image part = make_image(w, im.h, im.c);
+ int r, c, k;
+ float w_scale = (float)(im.w - 1) / (w - 1);
+ float h_scale = (float)(im.h - 1) / (h - 1);
+ for(k = 0; k < im.c; ++k){
+ for(r = 0; r < im.h; ++r){
+ for(c = 0; c < w; ++c){
+ float val = 0;
+ if(c == w-1){
+ val = get_pixel(im, im.w-1, r, k);
+ } else {
+ float sx = c*w_scale;
+ int ix = (int) sx;
+ float dx = sx - ix;
+ val = (1 - dx) * get_pixel(im, ix, r, k) + dx * get_pixel(im, ix+1, r, k);
}
+ set_pixel(part, c, r, k, val);
}
}
- for(k = 0; k < im.c; ++k){
- for(r = 0; r < h; ++r){
- float sy = r*h_scale;
- int iy = (int) sy;
- float dy = sy - iy;
- for(c = 0; c < w; ++c){
- float val = (1-dy) * get_pixel(part, c, iy, k);
- set_pixel(resized, c, r, k, val);
- }
- if(r == h-1) continue;
- for(c = 0; c < w; ++c){
- float val = dy * get_pixel(part, c, iy+1, k);
- add_pixel(resized, c, r, k, val);
- }
+ }
+ for(k = 0; k < im.c; ++k){
+ for(r = 0; r < h; ++r){
+ float sy = r*h_scale;
+ int iy = (int) sy;
+ float dy = sy - iy;
+ for(c = 0; c < w; ++c){
+ float val = (1-dy) * get_pixel(part, c, iy, k);
+ set_pixel(resized, c, r, k, val);
+ }
+ if(r == h-1) continue;
+ for(c = 0; c < w; ++c){
+ float val = dy * get_pixel(part, c, iy+1, k);
+ add_pixel(resized, c, r, k, val);
}
}
-
- free_image(part);
- return resized;
}
- void test_resize(char *filename)
- {
- image im = load_image(filename, 0,0, 3);
- image gray = grayscale_image(im);
+ free_image(part);
+ return resized;
+}
+
+void test_resize(char *filename)
+{
+ image im = load_image(filename, 0,0, 3);
+ image gray = grayscale_image(im);
- image sat2 = copy_image(im);
- saturate_image(sat2, 2);
+ image sat2 = copy_image(im);
+ saturate_image(sat2, 2);
- image sat5 = copy_image(im);
- saturate_image(sat5, .5);
+ image sat5 = copy_image(im);
+ saturate_image(sat5, .5);
- image exp2 = copy_image(im);
- exposure_image(exp2, 2);
+ image exp2 = copy_image(im);
+ exposure_image(exp2, 2);
- image exp5 = copy_image(im);
- exposure_image(exp5, .5);
+ image exp5 = copy_image(im);
+ exposure_image(exp5, .5);
- show_image(im, "Original");
- show_image(gray, "Gray");
- show_image(sat2, "Saturation-2");
- show_image(sat5, "Saturation-.5");
- show_image(exp2, "Exposure-2");
- show_image(exp5, "Exposure-.5");
+ show_image(im, "Original");
+ show_image(gray, "Gray");
+ show_image(sat2, "Saturation-2");
+ show_image(sat5, "Saturation-.5");
+ show_image(exp2, "Exposure-2");
+ show_image(exp5, "Exposure-.5");
#ifdef OPENCV
- cvWaitKey(0);
+ cvWaitKey(0);
#endif
- }
+}
#ifdef OPENCV
- image ipl_to_image(IplImage* src)
- {
- unsigned char *data = (unsigned char *)src->imageData;
- int h = src->height;
- int w = src->width;
- int c = src->nChannels;
- int step = src->widthStep;
- image out = make_image(w, h, c);
- int i, j, k, count=0;;
-
- for(k= 0; k < c; ++k){
- for(i = 0; i < h; ++i){
- for(j = 0; j < w; ++j){
- out.data[count++] = data[i*step + j*c + k]/255.;
- }
+image ipl_to_image(IplImage* src)
+{
+ unsigned char *data = (unsigned char *)src->imageData;
+ int h = src->height;
+ int w = src->width;
+ int c = src->nChannels;
+ int step = src->widthStep;
+ image out = make_image(w, h, c);
+ int i, j, k, count=0;;
+
+ for(k= 0; k < c; ++k){
+ for(i = 0; i < h; ++i){
+ for(j = 0; j < w; ++j){
+ out.data[count++] = data[i*step + j*c + k]/255.;
}
}
- return out;
}
+ return out;
+}
- image load_image_cv(char *filename, int channels)
- {
- IplImage* src = 0;
- int flag = -1;
- if (channels == 0) flag = -1;
- else if (channels == 1) flag = 0;
- else if (channels == 3) flag = 1;
- else {
- fprintf(stderr, "OpenCV can't force load with %d channels\n", channels);
- }
+image load_image_cv(char *filename, int channels)
+{
+ IplImage* src = 0;
+ int flag = -1;
+ if (channels == 0) flag = -1;
+ else if (channels == 1) flag = 0;
+ else if (channels == 3) flag = 1;
+ else {
+ fprintf(stderr, "OpenCV can't force load with %d channels\n", channels);
+ }
- if( (src = cvLoadImage(filename, flag)) == 0 )
- {
- printf("Cannot load file image %s\n", filename);
- exit(0);
- }
- image out = ipl_to_image(src);
- cvReleaseImage(&src);
- rgbgr_image(out);
- return out;
+ if( (src = cvLoadImage(filename, flag)) == 0 )
+ {
+ printf("Cannot load file image %s\n", filename);
+ exit(0);
}
+ image out = ipl_to_image(src);
+ cvReleaseImage(&src);
+ rgbgr_image(out);
+ return out;
+}
#endif
- image load_image_stb(char *filename, int channels)
- {
- int w, h, c;
- unsigned char *data = stbi_load(filename, &w, &h, &c, channels);
- if (!data) {
- fprintf(stderr, "Cannot load file image %s\nSTB Reason: %s\n", filename, stbi_failure_reason());
- exit(0);
- }
- if(channels) c = channels;
- int i,j,k;
- image im = make_image(w, h, c);
- for(k = 0; k < c; ++k){
- for(j = 0; j < h; ++j){
- for(i = 0; i < w; ++i){
- int dst_index = i + w*j + w*h*k;
- int src_index = k + c*i + c*w*j;
- im.data[dst_index] = (float)data[src_index]/255.;
- }
+image load_image_stb(char *filename, int channels)
+{
+ int w, h, c;
+ unsigned char *data = stbi_load(filename, &w, &h, &c, channels);
+ if (!data) {
+ fprintf(stderr, "Cannot load file image %s\nSTB Reason: %s\n", filename, stbi_failure_reason());
+ exit(0);
+ }
+ if(channels) c = channels;
+ int i,j,k;
+ image im = make_image(w, h, c);
+ for(k = 0; k < c; ++k){
+ for(j = 0; j < h; ++j){
+ for(i = 0; i < w; ++i){
+ int dst_index = i + w*j + w*h*k;
+ int src_index = k + c*i + c*w*j;
+ im.data[dst_index] = (float)data[src_index]/255.;
}
}
- free(data);
- return im;
}
+ free(data);
+ return im;
+}
- image load_image(char *filename, int w, int h, int c)
- {
+image load_image(char *filename, int w, int h, int c)
+{
#ifdef OPENCV
- image out = load_image_cv(filename, c);
+ image out = load_image_cv(filename, c);
#else
- image out = load_image_stb(filename, c);
+ image out = load_image_stb(filename, c);
#endif
- if((h && w) && (h != out.h || w != out.w)){
- image resized = resize_image(out, w, h);
- free_image(out);
- out = resized;
- }
- return out;
+ if((h && w) && (h != out.h || w != out.w)){
+ image resized = resize_image(out, w, h);
+ free_image(out);
+ out = resized;
}
+ return out;
+}
- image load_image_color(char *filename, int w, int h)
- {
- return load_image(filename, w, h, 3);
- }
+image load_image_color(char *filename, int w, int h)
+{
+ return load_image(filename, w, h, 3);
+}
- image get_image_layer(image m, int l)
- {
- image out = make_image(m.w, m.h, 1);
- int i;
- for(i = 0; i < m.h*m.w; ++i){
- out.data[i] = m.data[i+l*m.h*m.w];
- }
- return out;
+image get_image_layer(image m, int l)
+{
+ image out = make_image(m.w, m.h, 1);
+ int i;
+ for(i = 0; i < m.h*m.w; ++i){
+ out.data[i] = m.data[i+l*m.h*m.w];
}
+ return out;
+}
- float get_pixel(image m, int x, int y, int c)
- {
- assert(x < m.w && y < m.h && c < m.c);
- return m.data[c*m.h*m.w + y*m.w + x];
- }
- float get_pixel_extend(image m, int x, int y, int c)
- {
- if(x < 0 || x >= m.w || y < 0 || y >= m.h || c < 0 || c >= m.c) return 0;
- return get_pixel(m, x, y, c);
- }
- void set_pixel(image m, int x, int y, int c, float val)
- {
- assert(x < m.w && y < m.h && c < m.c);
- m.data[c*m.h*m.w + y*m.w + x] = val;
- }
- void add_pixel(image m, int x, int y, int c, float val)
- {
- assert(x < m.w && y < m.h && c < m.c);
- m.data[c*m.h*m.w + y*m.w + x] += val;
- }
+float get_pixel(image m, int x, int y, int c)
+{
+ assert(x < m.w && y < m.h && c < m.c);
+ return m.data[c*m.h*m.w + y*m.w + x];
+}
+float get_pixel_extend(image m, int x, int y, int c)
+{
+ if(x < 0 || x >= m.w || y < 0 || y >= m.h || c < 0 || c >= m.c) return 0;
+ return get_pixel(m, x, y, c);
+}
+void set_pixel(image m, int x, int y, int c, float val)
+{
+ assert(x < m.w && y < m.h && c < m.c);
+ m.data[c*m.h*m.w + y*m.w + x] = val;
+}
+void add_pixel(image m, int x, int y, int c, float val)
+{
+ assert(x < m.w && y < m.h && c < m.c);
+ m.data[c*m.h*m.w + y*m.w + x] += val;
+}
- void print_image(image m)
- {
- int i, j, k;
- for(i =0 ; i < m.c; ++i){
- for(j =0 ; j < m.h; ++j){
- for(k = 0; k < m.w; ++k){
- printf("%.2lf, ", m.data[i*m.h*m.w + j*m.w + k]);
- if(k > 30) break;
- }
- printf("\n");
- if(j > 30) break;
+void print_image(image m)
+{
+ int i, j, k;
+ for(i =0 ; i < m.c; ++i){
+ for(j =0 ; j < m.h; ++j){
+ for(k = 0; k < m.w; ++k){
+ printf("%.2lf, ", m.data[i*m.h*m.w + j*m.w + k]);
+ if(k > 30) break;
}
printf("\n");
+ if(j > 30) break;
}
printf("\n");
}
+ printf("\n");
+}
- image collapse_images_vert(image *ims, int n)
- {
- int color = 1;
- int border = 1;
- int h,w,c;
- w = ims[0].w;
- h = (ims[0].h + border) * n - border;
- c = ims[0].c;
- if(c != 3 || !color){
- w = (w+border)*c - border;
- c = 1;
- }
-
- image filters = make_image(w, h, c);
- int i,j;
- for(i = 0; i < n; ++i){
- int h_offset = i*(ims[0].h+border);
- image copy = copy_image(ims[i]);
- //normalize_image(copy);
- if(c == 3 && color){
- embed_image(copy, filters, 0, h_offset);
- }
- else{
- for(j = 0; j < copy.c; ++j){
- int w_offset = j*(ims[0].w+border);
- image layer = get_image_layer(copy, j);
- embed_image(layer, filters, w_offset, h_offset);
- free_image(layer);
- }
+image collapse_images_vert(image *ims, int n)
+{
+ int color = 1;
+ int border = 1;
+ int h,w,c;
+ w = ims[0].w;
+ h = (ims[0].h + border) * n - border;
+ c = ims[0].c;
+ if(c != 3 || !color){
+ w = (w+border)*c - border;
+ c = 1;
+ }
+
+ image filters = make_image(w, h, c);
+ int i,j;
+ for(i = 0; i < n; ++i){
+ int h_offset = i*(ims[0].h+border);
+ image copy = copy_image(ims[i]);
+ //normalize_image(copy);
+ if(c == 3 && color){
+ embed_image(copy, filters, 0, h_offset);
+ }
+ else{
+ for(j = 0; j < copy.c; ++j){
+ int w_offset = j*(ims[0].w+border);
+ image layer = get_image_layer(copy, j);
+ embed_image(layer, filters, w_offset, h_offset);
+ free_image(layer);
}
- free_image(copy);
- }
- return filters;
- }
-
- image collapse_images_horz(image *ims, int n)
- {
- int color = 1;
- int border = 1;
- int h,w,c;
- int size = ims[0].h;
- h = size;
- w = (ims[0].w + border) * n - border;
- c = ims[0].c;
- if(c != 3 || !color){
- h = (h+border)*c - border;
- c = 1;
}
+ free_image(copy);
+ }
+ return filters;
+}
- image filters = make_image(w, h, c);
- int i,j;
- for(i = 0; i < n; ++i){
- int w_offset = i*(size+border);
- image copy = copy_image(ims[i]);
- //normalize_image(copy);
- if(c == 3 && color){
- embed_image(copy, filters, w_offset, 0);
- }
- else{
- for(j = 0; j < copy.c; ++j){
- int h_offset = j*(size+border);
- image layer = get_image_layer(copy, j);
- embed_image(layer, filters, w_offset, h_offset);
- free_image(layer);
- }
+image collapse_images_horz(image *ims, int n)
+{
+ int color = 1;
+ int border = 1;
+ int h,w,c;
+ int size = ims[0].h;
+ h = size;
+ w = (ims[0].w + border) * n - border;
+ c = ims[0].c;
+ if(c != 3 || !color){
+ h = (h+border)*c - border;
+ c = 1;
+ }
+
+ image filters = make_image(w, h, c);
+ int i,j;
+ for(i = 0; i < n; ++i){
+ int w_offset = i*(size+border);
+ image copy = copy_image(ims[i]);
+ //normalize_image(copy);
+ if(c == 3 && color){
+ embed_image(copy, filters, w_offset, 0);
+ }
+ else{
+ for(j = 0; j < copy.c; ++j){
+ int h_offset = j*(size+border);
+ image layer = get_image_layer(copy, j);
+ embed_image(layer, filters, w_offset, h_offset);
+ free_image(layer);
}
- free_image(copy);
- }
- return filters;
- }
-
- void show_images(image *ims, int n, char *window)
- {
- image m = collapse_images_vert(ims, n);
- /*
- int w = 448;
- int h = ((float)m.h/m.w) * 448;
- if(h > 896){
- h = 896;
- w = ((float)m.w/m.h) * 896;
}
- image sized = resize_image(m, w, h);
- */
- normalize_image(m);
- image sized = resize_image(m, m.w, m.h);
- save_image(sized, window);
- show_image(sized, window);
- free_image(sized);
- free_image(m);
+ free_image(copy);
}
+ return filters;
+}
- void free_image(image m)
- {
- free(m.data);
- }
+void show_images(image *ims, int n, char *window)
+{
+ image m = collapse_images_vert(ims, n);
+ /*
+ int w = 448;
+ int h = ((float)m.h/m.w) * 448;
+ if(h > 896){
+ h = 896;
+ w = ((float)m.w/m.h) * 896;
+ }
+ image sized = resize_image(m, w, h);
+ */
+ normalize_image(m);
+ image sized = resize_image(m, m.w, m.h);
+ save_image(sized, window);
+ show_image(sized, window);
+ free_image(sized);
+ free_image(m);
+}
+
+void free_image(image m)
+{
+ free(m.data);
+}
diff --git a/src/region_layer.c b/src/region_layer.c
index ecb89c64bda9dc7f339fc777a5fde4f3e2e24234..39af5ee8c528ecf795eac1a28e43d87f5bec89e4 100644
--- a/src/region_layer.c
+++ b/src/region_layer.c
@@ -61,6 +61,7 @@ void forward_region_layer(const region_layer l, network_state state)
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;
@@ -90,6 +91,7 @@ void forward_region_layer(const region_layer l, network_state state)
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);
@@ -151,7 +153,7 @@ void forward_region_layer(const region_layer l, network_state state)
LOGISTIC, l.delta + index + locations*l.classes);
}
}
- printf("Region Avg IOU: %f, Avg Cat Pred: %f, Avg Obj: %f, Avg Any: %f, count: %d\n", avg_iou/count, avg_cat/count, avg_obj/count, avg_anyobj/(l.batch*locations*l.n), count);
+ printf("Region Avg IOU: %f, Pos Cat: %f, All Cat: %f, Pos Obj: %f, Any Obj: %f, count: %d\n", avg_iou/count, avg_cat/count, avg_allcat/(count*l.classes), avg_obj/count, avg_anyobj/(l.batch*locations*l.n), count);
}
}
diff --git a/src/swag.c b/src/swag.c
index 1398192de17fe489af090f52391cad6a86987c05..37dde36a224a38766b3e1bb25965416d205d178d 100644
--- a/src/swag.c
+++ b/src/swag.c
@@ -132,21 +132,22 @@ void train_swag(char *cfgfile, char *weightfile)
void convert_swag_detections(float *predictions, int classes, int num, int square, int side, int w, int h, float thresh, float **probs, box *boxes)
{
int i,j,n;
- int per_cell = 5*num+classes;
+ //int per_cell = 5*num+classes;
for (i = 0; i < side*side; ++i){
int row = i / side;
int col = i % side;
for(n = 0; n < num; ++n){
- int offset = i*per_cell + 5*n;
- float scale = predictions[offset];
int index = i*num + n;
- boxes[index].x = (predictions[offset + 1] + col) / side * w;
- boxes[index].y = (predictions[offset + 2] + row) / side * h;
- boxes[index].w = pow(predictions[offset + 3], (square?2:1)) * w;
- boxes[index].h = pow(predictions[offset + 4], (square?2:1)) * h;
+ int p_index = side*side*classes + i*num + n;
+ float scale = predictions[p_index];
+ int box_index = side*side*(classes + num) + (i*num + n)*4;
+ boxes[index].x = (predictions[box_index + 0] + col) / side * w;
+ boxes[index].y = (predictions[box_index + 1] + row) / side * h;
+ boxes[index].w = pow(predictions[box_index + 2], (square?2:1)) * w;
+ boxes[index].h = pow(predictions[box_index + 3], (square?2:1)) * h;
for(j = 0; j < classes; ++j){
- offset = i*per_cell + 5*num;
- float prob = scale*predictions[offset+j];
+ int class_index = i*classes;
+ float prob = scale*predictions[class_index+j];
probs[index][j] = (prob > thresh) ? prob : 0;
}
}
diff --git a/src/writing.c b/src/writing.c
index cfbc5fd16dcae67919ddc2f430e02096053ce6bf..e0c1d1a83cb0d6c64b99916e5c77474aaa936a0b 100644
--- a/src/writing.c
+++ b/src/writing.c
@@ -2,8 +2,13 @@
#include "utils.h"
#include "parser.h"
+#ifdef OPENCV
+#include "opencv2/highgui/highgui_c.h"
+#endif
+
void train_writing(char *cfgfile, char *weightfile)
{
+ char *backup_directory = "/home/pjreddie/backup/";
data_seed = time(0);
srand(time(0));
float avg_loss = -1;
@@ -23,41 +28,78 @@ void train_writing(char *cfgfile, char *weightfile)
while(1){
++i;
time=clock();
- data train = load_data_writing(paths, imgs, plist->size, 512, 512);
+ data train = load_data_writing(paths, imgs, plist->size, 256, 256, 1);
+ printf("Loaded %lf seconds\n",sec(clock()-time));
+ time=clock();
float loss = train_network(net, train);
- #ifdef GPU
- float *out = get_network_output_gpu(net);
- #else
- float *out = get_network_output(net);
- #endif
- image pred = float_to_image(64, 64, 1, out);
- print_image(pred);
-/*
- image im = float_to_image(256, 256, 3, train.X.vals[0]);
- image lab = float_to_image(64, 64, 1, train.y.vals[0]);
+ /*
image pred = float_to_image(64, 64, 1, out);
- show_image(im, "image");
- show_image(lab, "label");
- print_image(lab);
- show_image(pred, "pred");
- cvWaitKey(0);
+ print_image(pred);
*/
+ /*
+ image im = float_to_image(256, 256, 3, train.X.vals[0]);
+ image lab = float_to_image(64, 64, 1, train.y.vals[0]);
+ image pred = float_to_image(64, 64, 1, out);
+ show_image(im, "image");
+ show_image(lab, "label");
+ print_image(lab);
+ show_image(pred, "pred");
+ cvWaitKey(0);
+ */
+
if(avg_loss == -1) avg_loss = loss;
avg_loss = avg_loss*.9 + loss*.1;
printf("%d: %f, %f avg, %lf seconds, %d images\n", i, loss, avg_loss, sec(clock()-time), *net.seen);
free_data(train);
- if((i % 20000) == 0) net.learning_rate *= .1;
//if(i%100 == 0 && net.learning_rate > .00001) net.learning_rate *= .97;
if(i%1000==0){
char buff[256];
- sprintf(buff, "/home/pjreddie/imagenet_backup/%s_%d.weights",base, i);
+ sprintf(buff, "%s/%s_%d.weights", backup_directory, base, i);
save_weights(net, buff);
}
}
}
+void test_writing(char *cfgfile, char *weightfile, char *outfile)
+{
+ network net = parse_network_cfg(cfgfile);
+ if(weightfile){
+ load_weights(&net, weightfile);
+ }
+ set_batch_network(&net, 1);
+ srand(2222222);
+ clock_t time;
+ char filename[256];
+
+ fgets(filename, 256, stdin);
+ strtok(filename, "\n");
+ image im = load_image_color(filename, 0, 0);
+ //image im = load_image_color("/home/pjreddie/darknet/data/figs/C02-1001-Figure-1.png", 0, 0);
+ image sized = resize_image(im, net.w, net.h);
+ printf("%d %d %d\n", im.h, im.w, im.c);
+ float *X = sized.data;
+ time=clock();
+ network_predict(net, X);
+ printf("%s: Predicted in %f seconds.\n", filename, sec(clock()-time));
+ image pred = get_network_image(net);
+
+ if (outfile) {
+ printf("Save image as %s.png (shape: %d %d)\n", outfile, pred.w, pred.h);
+ save_image(pred, outfile);
+ } else {
+ show_image(pred, "prediction");
+#ifdef OPENCV
+ cvWaitKey(0);
+ cvDestroyAllWindows();
+#endif
+ }
+
+ free_image(im);
+ free_image(sized);
+}
+
void run_writing(int argc, char **argv)
{
if(argc < 4){
@@ -67,6 +109,8 @@ void run_writing(int argc, char **argv)
char *cfg = argv[3];
char *weights = (argc > 4) ? argv[4] : 0;
+ char *outfile = (argc > 5) ? argv[5] : 0;
if(0==strcmp(argv[2], "train")) train_writing(cfg, weights);
+ else if(0==strcmp(argv[2], "test")) test_writing(cfg, weights, outfile);
}