Darknet: Crop object detected

[halimB8]

hello ,
How can I use the function "crop_image()" inside the function "draw_detections_cv_v3()" to crop all the object detected on a video and save them in a folder

[AlexeyAB]

@halimB8 Hi,

1. Just un-comment these lines: https://github.com/AlexeyAB/darknet/blob/a720bf74e233c956e5281871ede722a2c7cf54db/src/image.c#L520-L534

2. Recompile Darknet

3. Create result_img directory near with darknet-executable file
4. Run darknet on the video file:
   ./darknet detector demo data/coco.data cfg/yolov3.cfg yolov3.weights test.mp4
5. Look at the result_img directory
   You will see files like img_<frame_id>_<img_id>_<class_id>.jpg with cropped detected objects.

About it: https://github.com/AlexeyAB/darknet/issues/846#issuecomment-390785676

[halimB8]

it works
thank you so much :)

[AlexeyAB]

@james203198191 What do you mean? What error did you get?

[AlexeyAB]

@james203198191

1. Did you uncomment these lines? https://github.com/AlexeyAB/darknet/blob/a720bf74e233c956e5281871ede722a2c7cf54db/src/image.c#L520-L534

1. Did you create this folder result_img ?

2. Did you run exactly this command?
   ./darknet detector demo data/coco.data cfg/yolov3.cfg yolov3.weights test.mp4

And you didn't get images img_<frame_id>_<img_id>_<class_id>.jpg in the result_img folder?

[james203198191]

i solved it
thank you for your reply

[dfsaw]

I uncommented the above lines. However folder result_img is empty. I ran following command
./darknet detector test data/obj.data cfg/yolov3-tiny-obj.cfg /backup/yolov3-tiny.weights 1.jpeg

[AlexeyAB]

@dfsaw
Currently it is implemented for video only (not for images):
./darknet detector demo data/coco.data cfg/yolov3.cfg yolov3.weights test.mp4

[dfsaw]

@AlexeyAB okay. is there any way i can get coordinates of all detected objects in a text file

[AlexeyAB]

@dfsaw
Yes: https://github.com/AlexeyAB/darknet/issues/1099#issuecomment-400501945

[rachana88]

"./darknet detector demo data/coco.data cfg/yolov3.cfg yolov3.weights test.mp4 " for the above command the coco.data file is in cfg folder and not data.Also i am gettting error of demo needs opencv for webcam images

[AlexeyAB]

@rachana88
Yes, use cfg/coco.data instead of data/coco.data. And you should have your own video file test.mp4
./darknet detector demo cfg/coco.data cfg/yolov3.cfg yolov3.weights test.mp4

[rachana88]

ok, thanks!!

But Getting the same error of "

Demo needs OpenCV for webcam images

On Thu, Sep 13, 2018 at 2:05 PM Alexey [e-mail] wrote:

@rachana88 https://github.com/rachana88
Yes, use cfg/coco.data instead of data/coco.data. And you should have your
own video file test.mp4
./darknet detector demo cfg/coco.data cfg/yolov3.cfg yolov3.weights
test.mp4

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/AlexeyAB/darknet/issues/954#issuecomment-421152916,
or mute the thread
https://github.com/notifications/unsubscribe-auth/Ai-bMQltJT0FzpcC2MY4pgkYc58PwmrRks5uasiLgaJpZM4UVaaI
.

--
Rachana R. Desai
Data scientist
Cisco

[rachana88]

ok, thanks!!

But Getting the same error of "
Demo needs OpenCV for webcam images

[AlexeyAB]

@rachana88
Yes, you should compile Darknet with OPENCV=1 in the Makefile. For this OpenCV should be installed: https://opencv.org/releases.html
OpenCV 3.3.0 or lower.

Without OpenCV you can't detect on video.

[rachana88]

SO, i am sorry but now getting the error of "(Demo:392): Gtk-WARNING **: cannot open display:
"

[halimB8]

add "-dont-show" in the end of your command to don't display the video

[rachana88]

Hello,
Sorry but having the same error!!

(Demo:805): Gtk-WARNING **: cannot open display:

On Fri, Sep 14, 2018 at 2:59 AM halimB8 [e-mail] wrote:

add "-dont-show" in the end of your command to don't display the video

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/AlexeyAB/darknet/issues/954#issuecomment-421302396,
or mute the thread
https://github.com/notifications/unsubscribe-auth/Ai-bMXuR1qqqetIslZKWGZ7O1aby87Ffks5ua334gaJpZM4UVaaI
.

--
Rachana R. Desai
Data scientist
Cisco

[nwthilina]

How can i crop yolo annotated bounding box result image. is there any way to do that

[nwthilina]

How can i crop yolo annotated bounding box result image. is there any way to do that

Found the Solution
add these line after 448

             image c1 = crop_image(im, left, top, right-left, bot-top);
         save_image(c1, "99");

[Rahul-Venugopal]

@AlexeyAB @nwthilina I made the changes as u suggested and now the detected images are saved in results_img folder. Thanks for the option.
Is it possible to only save the images on the basis of a flag ? Now it saves whenever we run a video and it would be really great to save detected images only when activated with a flag.
is there any option for that ? If not , is it possible achieve this only by editing image.c accordingly ?

[sabrashahid0]

@AlexeyAB
Video-stream stoped!
Stream closed.
Video-stream stoped!
Stream closed.
Segmentation fault (core dumped)
this is the error i am getting...help needed!
i followed all the steps you said.

[mnalam30]

How about the crop object detected in python?

[MuhammadAsadJaved]

How can i crop yolo annotated bounding box result image. is there any way to do that

Found the Solution
add these line after 448

             image c1 = crop_image(im, left, top, right-left, bot-top);
       save_image(c1, "99");

Hello , I also have the same task, can you guide, please? I am using a single image and single object detection (LCD detection and cropping)

[MuhammadAsadJaved]

How can i crop yolo annotated bounding box result image. is there any way to do that

Found the Solution
add these line after 448

             image c1 = crop_image(im, left, top, right-left, bot-top);
         save_image(c1, "99");

Hello , I also have the same task, can you guide, please? I am using a single image and single object detection (LCD detection and cropping)

In which file i should wite these two lines?

[nwthilina]

U should code in image.c file

On Tue, 18 Jun 2019, 15:51 MuhammadAsadJaved, [e-mail]
wrote:

How can i crop yolo annotated bounding box result image. is there any way
to do that

Found the Solution
add these line after 448

         image c1 = crop_image(im, left, top, right-left, bot-top);
   save_image(c1, "99");

Hello , I also have the same task, can you guide, please? I am using a
single image and single object detection (LCD detection and cropping)

In which file i should wite these two lines?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/AlexeyAB/darknet/issues/954?email_source=notifications&email_token=AECBIXMMMUEC4RMTNFW2CWDP3CZKXA5CNFSM4FCVU2EKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODX55NVY#issuecomment-503043799,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AECBIXND7AIBELX5FBIAJFTP3CZKXANCNFSM4FCVU2EA
.

[MuhammadAsadJaved]

@nwthilina
Sir my code differ from your's, maybe it's updated version. I have this code at line 448, Can you send me the function name and some lines of code from function after which I should put these two lines? I have tried these lines at the end of this function. also after line 448, but it still show the same output.

void draw_detections(image im, int num, float thresh, box boxes, float *probs, char *names, image *alphabet, int classes)
{
int i;

for(i = 0; i < num; ++i){
    int class_id = max_index(probs[i], classes);
    float prob = probs[i][class_id];
    if(prob > thresh){

        //// for comparison with OpenCV version of DNN Darknet Yolo v2
        //printf("\n %f, %f, %f, %f, ", boxes[i].x, boxes[i].y, boxes[i].w, boxes[i].h);
        // int k;
        //for (k = 0; k < classes; ++k) {
        //    printf("%f, ", probs[i][k]);
        //}
        //printf("\n");

448 line>> int width = im.h * .012;

        if(0){
            width = pow(prob, 1./2.)*10+1;
            alphabet = 0;
        }

        int offset = class_id*123457 % classes;
        float red = get_color(2,offset,classes);
        float green = get_color(1,offset,classes);
        float blue = get_color(0,offset,classes);
        float rgb[3];

        //width = prob*20+2;

        rgb[0] = red;
        rgb[1] = green;
        rgb[2] = blue;
        box b = boxes[i];

        int left  = (b.x-b.w/2.)*im.w;
        int right = (b.x+b.w/2.)*im.w;
        int top   = (b.y-b.h/2.)*im.h;
        int bot   = (b.y+b.h/2.)*im.h;

        if(left < 0) left = 0;
        if(right > im.w-1) right = im.w-1;
        if(top < 0) top = 0;
        if(bot > im.h-1) bot = im.h-1;
        printf("%s: %.0f%%", names[class_id], prob * 100);

        //printf(" - id: %d, x_center: %d, y_center: %d, width: %d, height: %d",
        //    class_id, (right + left) / 2, (bot - top) / 2, right - left, bot - top);

        printf("\n");
        draw_box_width(im, left, top, right, bot, width, red, green, blue);
        if (alphabet) {
            image label = get_label(alphabet, names[class_id], (im.h*.03)/10);
            draw_label(im, top + width, left, label, rgb);
            image c1 = crop_image(im, left, top, right-left, bot-top);
            save_image(c1, "99");
        }
    }
}

}

[nwthilina]

void draw_detections_v3(image im, detection dets, int num, float thresh,
char *names, image *alphabet, int classes, int ext_output)
{
int selected_detections_num;
detection_with_class selected_detections = get_actual_detections(dets,
num, thresh, &selected_detections_num);

// text output
qsort(selected_detections, selected_detections_num,

sizeof(selected_detections), compare_by_lefts);
int i;
for (i = 0; i < selected_detections_num; ++i) {
const int best_class = selected_detections[i].best_class;
printf("%s: %.0f%%", names[best_class],
selected_detections[i].det.prob[best_class] * 100);
if (ext_output)
printf("\t(left_x: %4.0f top_y: %4.0f width: %4.0f
height: %4.0f)\n",
(selected_detections[i].det.bbox.x -
selected_detections[i].det.bbox.w / 2)im.w,
(selected_detections[i].det.bbox.y -
selected_detections[i].det.bbox.h / 2)im.h,
selected_detections[i].det.bbox.wim.w,
selected_detections[i].det.bbox.h*im.h);
else
printf("\n");
int j;
for (j = 0; j < classes; ++j) {
if (selected_detections[i].det.prob[j] > thresh && j !=
best_class) {
printf("%s: %.0f%%\n", names[j],
selected_detections[i].det.prob[j] * 100);
}
}
}

// image output
qsort(selected_detections, selected_detections_num,

sizeof(*selected_detections), compare_by_probs);
for (i = 0; i < selected_detections_num; ++i) {
int width = im.h * .006;
if (width < 1)
width = 1;

        /*
        if(0){
        width = pow(prob, 1./2.)*10+1;
        alphabet = 0;
        }
        */

        //printf("%d %s: %.0f%%\n", i,

names[selected_detections[i].best_class], prob*100);
int offset = selected_detections[i].best_class * 123457 %
classes;
float red = get_color(2, offset, classes);
float green = get_color(1, offset, classes);
float blue = get_color(0, offset, classes);
float rgb[3];

        //width = prob*20+2;

        rgb[0] = red;
        rgb[1] = green;
        rgb[2] = blue;
        box b = selected_detections[i].det.bbox;
        printf("%f %f %f %f\n", b.x, b.y, b.w, b.h);

        int left = (b.x - b.w / 2.)*im.w;
        int right = (b.x + b.w / 2.)*im.w;
        int top = (b.y - b.h / 2.)*im.h;
        int bot = (b.y + b.h / 2.)*im.h;

        if (left < 0) left = 0;
        if (right > im.w - 1) right = im.w - 1;
        if (top < 0) top = 0;
        if (bot > im.h - 1) bot = im.h - 1;

        //int b_x_center = (left + right) / 2;
        //int b_y_center = (top + bot) / 2;
        //int b_width = right - left;
        //int b_height = bot - top;
        //sprintf(labelstr, "%d x %d - w: %d, h: %d", b_x_center,

b_y_center, b_width, b_height);

        printf("Cropping Vehicle Plate");
        draw_box_width(im, left, top, right, bot, width, red, green,

blue);
image c1 = crop_image(im, left, top, right-left, bot-top);
save_image(c1, "crop_plate");

         if (alphabet) {
            char labelstr[4096] = { 0 };
            strcat(labelstr, names[selected_detections[i].best_class]);
            int j;
            for (j = 0; j < classes; ++j) {
                if (selected_detections[i].det.prob[j] > thresh && j !=

selected_detections[i].best_class) {
strcat(labelstr, ", ");
strcat(labelstr, names[j]);
}
}
image label = get_label_v3(alphabet, labelstr, (im.h.03));
draw_label(im, top + width, left, label, rgb);
free_image(label);
}
if (selected_detections[i].det.mask) {
image mask = float_to_image(14, 14, 1,
selected_detections[i].det.mask);
image resized_mask = resize_image(mask, b.wim.w, b.h*im.h);
image tmask = threshold_image(resized_mask, .5);
embed_image(tmask, im, left, top);
free_image(mask);
free_image(resized_mask);
free_image(tmask);
}
}
free(selected_detections);
}

or use attached image.c file

Thilina Nuwan Rathnayaka

Mob : (+094)773714989

On Wed, Jun 19, 2019 at 8:40 AM MuhammadAsadJaved [e-mail]
wrote:

@nwthilina https://github.com/nwthilina
Sir my code differ from your's, maybe it's updated version. I have this
code at line 448, Can you send me the function name and some lines of code
from function after which I should put these two lines? I have tried these
lines at the end of this function. also after line 448, but it still show
the same output.

void draw_detections(image im, int num, float thresh, box boxes, float
*probs, char *names, image *alphabet, int classes)
{
int i;

for(i = 0; i < num; ++i){
int class_id = max_index(probs[i], classes);
float prob = probs[i][class_id];
if(prob > thresh){

    //// for comparison with OpenCV version of DNN Darknet Yolo v2
    //printf("\n %f, %f, %f, %f, ", boxes[i].x, boxes[i].y, boxes[i].w, boxes[i].h);
    // int k;
    //for (k = 0; k < classes; ++k) {
    //    printf("%f, ", probs[i][k]);
    //}
    //printf("\n");

448 line>> int width = im.h * .012;

    if(0){
        width = pow(prob, 1./2.)*10+1;
        alphabet = 0;
    }

    int offset = class_id*123457 % classes;
    float red = get_color(2,offset,classes);
    float green = get_color(1,offset,classes);
    float blue = get_color(0,offset,classes);
    float rgb[3];

    //width = prob*20+2;

    rgb[0] = red;
    rgb[1] = green;
    rgb[2] = blue;
    box b = boxes[i];

    int left  = (b.x-b.w/2.)*im.w;
    int right = (b.x+b.w/2.)*im.w;
    int top   = (b.y-b.h/2.)*im.h;
    int bot   = (b.y+b.h/2.)*im.h;

    if(left < 0) left = 0;
    if(right > im.w-1) right = im.w-1;
    if(top < 0) top = 0;
    if(bot > im.h-1) bot = im.h-1;
    printf("%s: %.0f%%", names[class_id], prob * 100);

    //printf(" - id: %d, x_center: %d, y_center: %d, width: %d, height: %d",
    //    class_id, (right + left) / 2, (bot - top) / 2, right - left, bot - top);

    printf("\n");
    draw_box_width(im, left, top, right, bot, width, red, green, blue);
    if (alphabet) {
        image label = get_label(alphabet, names[class_id], (im.h*.03)/10);
        draw_label(im, top + width, left, label, rgb);
        image c1 = crop_image(im, left, top, right-left, bot-top);
        save_image(c1, "99");
    }
}

}

}

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/AlexeyAB/darknet/issues/954?email_source=notifications&email_token=AECBIXIHF3YJIPL63KYL2KLP3GPSNA5CNFSM4FCVU2EKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODYARSTI#issuecomment-503388493,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AECBIXOYHHI6R3BRAKCULP3P3GPSNANCNFSM4FCVU2EA
.

[thimabru1010]

@halimB8 Hi,

1. Just un-comment these lines: https://github.com/AlexeyAB/darknet/blob/a720bf74e233c956e5281871ede722a2c7cf54db/src/image.c#L520-L534
2. Recompile Darknet
3. Create result_img directory near with darknet-executable file
4. Run darknet on the video file:
   ./darknet detector demo data/coco.data cfg/yolov3.cfg yolov3.weights test.mp4
5. Look at the result_img directory
   You will see files like img_<frame_id>_<img_id>_<class_id>.jpg with cropped detected objects.

About it: #846 (comment)

I tried to recompile Darknet but I got the foloowing error:
In function ‘void draw_detections_cv_v3(mat_cv, detection, int, float, char, image, int, int)’:
./src/image_opencv.cpp:972:81: error: ‘class cv::Mat’ has no member named ‘width’
if(copy_img == NULL) copy_img = cvCreateImage(cvSize(show_img->width, show
^
compilation terminated due to -Wfatal-errors.
Makefile:149: recipe for target 'obj/image_opencv.o' failed
make: * [obj/image_opencv.o] Error 1

I already tried to solve but I was unsuccessful. It seems everyone could run after un-coment these lines. Any idea what's going on?

[vinicius-trev]

@AlexeyAB Does something changed from 2018 to now ?? I can't get the output images in result_img folder after recompile the code !!

[thimabru1010]

It works fine with the release from 7 June 2018: https://github.com/AlexeyAB/darknet/releases

However, it is still not working with the actual version using image_opencv.cpp

[hqq1990]

void draw_detections_v3(image im, detection dets, int num, float thresh, char names, image *alphabet, int classes, int ext_output) { int selected_detections_num; detection_with_class* selected_detections = get_actual_detections(dets, num, thresh, &selected_detections_num); // text output qsort(selected_detections, selected_detections_num, sizeof(selected_detections), compare_by_lefts); int i; for (i = 0; i < selected_detections_num; ++i) { const int best_class = selected_detections[i].best_class; printf("%s: %.0f%%", names[best_class], selected_detections[i].det.prob[best_class] * 100); if (ext_output) printf("\t(left_x: %4.0f top_y: %4.0f width: %4.0f height: %4.0f)\n", (selected_detections[i].det.bbox.x - selected_detections[i].det.bbox.w / 2)im.w, (selected_detections[i].det.bbox.y - selected_detections[i].det.bbox.h / 2)im.h, selected_detections[i].det.bbox.wim.w, selected_detections[i].det.bbox.him.h); else printf("\n"); int j; for (j = 0; j < classes; ++j) { if (selected_detections[i].det.prob[j] > thresh && j != best_class) { printf("%s: %.0f%%\n", names[j], selected_detections[i].det.prob[j] * 100); } } } // image output qsort(selected_detections, selected_detections_num, sizeof(selected_detections), compare_by_probs); for (i = 0; i < selected_detections_num; ++i) { int width = im.h * .006; if (width < 1) width = 1; /* if(0){ width = pow(prob, 1./2.)10+1; alphabet = 0; } */ //printf("%d %s: %.0f%%\n", i, names[selected_detections[i].best_class], prob100); int offset = selected_detections[i].best_class * 123457 % classes; float red = get_color(2, offset, classes); float green = get_color(1, offset, classes); float blue = get_color(0, offset, classes); float rgb[3]; //width = prob20+2; rgb[0] = red; rgb[1] = green; rgb[2] = blue; box b = selected_detections[i].det.bbox; printf("%f %f %f %f\n", b.x, b.y, b.w, b.h); int left = (b.x - b.w / 2.)im.w; int right = (b.x + b.w / 2.)im.w; int top = (b.y - b.h / 2.)im.h; int bot = (b.y + b.h / 2.)im.h; if (left < 0) left = 0; if (right > im.w - 1) right = im.w - 1; if (top < 0) top = 0; if (bot > im.h - 1) bot = im.h - 1; //int b_x_center = (left + right) / 2; //int b_y_center = (top + bot) / 2; //int b_width = right - left; //int b_height = bot - top; //sprintf(labelstr, "%d x %d - w: %d, h: %d", b_x_center, b_y_center, b_width, b_height); printf("Cropping Vehicle Plate"); draw_box_width(im, left, top, right, bot, width, red, green, blue); image c1 = crop_image(im, left, top, right-left, bot-top); save_image(c1, "crop_plate"); if (alphabet) { char labelstr[4096] = { 0 }; strcat(labelstr, names[selected_detections[i].best_class]); int j; for (j = 0; j < classes; ++j) { if (selected_detections[i].det.prob[j] > thresh && j != selected_detections[i].best_class) { strcat(labelstr, ", "); strcat(labelstr, names[j]); } } image label = get_label_v3(alphabet, labelstr, (im.h.03)); draw_label(im, top + width, left, label, rgb); free_image(label); } if (selected_detections[i].det.mask) { image mask = float_to_image(14, 14, 1, selected_detections[i].det.mask); image resized_mask = resize_image(mask, b.wim.w, b.him.h); image tmask = threshold_image(resized_mask, .5); embed_image(tmask, im, left, top); free_image(mask); free_image(resized_mask); free_image(tmask); } } free(selected_detections); } or use attached image.c file Thilina Nuwan Rathnayaka Mob : [phone]
…
On Wed, Jun 19, 2019 at 8:40 AM MuhammadAsadJaved @.> wrote: @nwthilina https://github.com/nwthilina Sir my code differ from your's, maybe it's updated version. I have this code at line 448, Can you send me the function name and some lines of code from function after which I should put these two lines? I have tried these lines at the end of this function. also after line 448, but it still show the same output. void draw_detections(image im, int num, float thresh, box *boxes, float *probs, char *names, image *alphabet, int classes) { int i; for(i = 0; i < num; ++i){ int class_id = max_index(probs[i], classes); float prob = probs[i][class_id]; if(prob > thresh){ //// for comparison with OpenCV version of DNN Darknet Yolo v2 //printf("\n %f, %f, %f, %f, ", boxes[i].x, boxes[i].y, boxes[i].w, boxes[i].h); // int k; //for (k = 0; k < classes; ++k) { // printf("%f, ", probs[i][k]); //} //printf("\n"); 448 line>> int width = im.h * .012; if(0){ width = pow(prob, 1./2.)10+1; alphabet = 0; } int offset = class_id123457 % classes; float red = get_color(2,offset,classes); float green = get_color(1,offset,classes); float blue = get_color(0,offset,classes); float rgb[3]; //width = prob20+2; rgb[0] = red; rgb[1] = green; rgb[2] = blue; box b = boxes[i]; int left = (b.x-b.w/2.)im.w; int right = (b.x+b.w/2.)im.w; int top = (b.y-b.h/2.)im.h; int bot = (b.y+b.h/2.)im.h; if(left < 0) left = 0; if(right > im.w-1) right = im.w-1; if(top < 0) top = 0; if(bot > im.h-1) bot = im.h-1; printf("%s: %.0f%%", names[class_id], prob * 100); //printf(" - id: %d, x_center: %d, y_center: %d, width: %d, height: %d", // class_id, (right + left) / 2, (bot - top) / 2, right - left, bot - top); printf("\n"); draw_box_width(im, left, top, right, bot, width, red, green, blue); if (alphabet) { image label = get_label(alphabet, names[class_id], (im.h.03)/10); draw_label(im, top + width, left, label, rgb); image c1 = crop_image(im, left, top, right-left, bot-top); save_image(c1, "99"); } } } } — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#954?email_source=notifications&email_token=AECBIXIHF3YJIPL63KYL2KLP3GPSNA5CNFSM4FCVU2EKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODYARSTI#issuecomment-503388493>, or mute the thread https://github.com/notifications/unsubscribe-auth/AECBIXOYHHI6R3BRAKCULP3P3GPSNANCNFSM4FCVU2EA .

dear sir:
i tried your code, but no result output. can you help me?

[Pulkitg64]

I un-comment following lines:
static int copied_frame_id = -1;
static image copy_img;
if (copied_frame_id != frame_id) {
copied_frame_id = frame_id;
if (copy_img.data) free_image(copy_img);
copy_img = copy_image(im);
}
image cropped_im = crop_image(copy_img, left, top, right - left, bot - top);
static int img_id = 0;
img_id++;
char image_name[1024];
int best_class_id = selected_detections[i].best_class;
sprintf(image_name, "result_img/img_%d_%d_%d_%s.jpg", frame_id, img_id, best_class_id, names[best_class_id]);
save_image(cropped_im, image_name);
free_image(cropped_im);

I also added following lines
image c1 = crop_image(im, left, top, right-left, bot-top);
save_image(c1, "99");

But I am getting no image in result_img folder for my videos but getting images when I am testing on single images. Please help me

[xinyee1997]

Facing the same problem too. Anyone can help?

[xinyee1997]

@AlexeyAB Can you help? I can't save the cropped image of detected object from video. It can only works in images.

[MuhammadAsadJaved]

Did you recompile project after changes ?

On Wed, Feb 12, 2020 at 9:39 PM xinyee1997 [e-mail] wrote:

@AlexeyAB https://github.com/AlexeyAB Can you help? I can't save the
cropped image of detected object from video. It can only works in images.

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
https://github.com/AlexeyAB/darknet/issues/954?email_source=notifications&email_token=AG4GR5E5HOBVMEA33GG4V7LRCP3Z3A5CNFSM4FCVU2EKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOELQZMUI#issuecomment-585209425,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/AG4GR5FNXONBBAKHTMP5WC3RCP3Z3ANCNFSM4FCVU2EA
.

[xinyee1997]

@MuhammadAsadJaved Yes. I faced the problem like @Pulkitg64 stated.
Are you able to crop detected objects from video? Can you guide us how to do so?

[MuhammadAsadJaved]

@xinyee1997 我还在图像上进行了测试，但没有在视频上进行尝试。

[Pulkitg64]

Did you recompile project after changes ?
…
On Wed, Feb 12, 2020 at 9:39 PM xinyee1997 @.*> wrote: @AlexeyAB https://github.com/AlexeyAB Can you help? I can't save the cropped image of detected object from video. It can only works in images. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#954?email_source=notifications&email_token=AG4GR5E5HOBVMEA33GG4V7LRCP3Z3A5CNFSM4FCVU2EKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOELQZMUI#issuecomment-585209425>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AG4GR5FNXONBBAKHTMP5WC3RCP3Z3ANCNFSM4FCVU2EA .

Yes I recompiled the project but works only for images

[qillbel]

@halimB8 Hi,

1. Just un-comment these lines: https://github.com/AlexeyAB/darknet/blob/a720bf74e233c956e5281871ede722a2c7cf54db/src/image.c#L520-L534
2. Recompile Darknet
3. Create result_img directory near with darknet-executable file
4. Run darknet on the video file:
   ./darknet detector demo data/coco.data cfg/yolov3.cfg yolov3.weights test.mp4
5. Look at the result_img directory
   You will see files like img_<frame_id>_<img_id>_<class_id>.jpg with cropped detected objects.

About it: #846 (comment)

I tried without success. I used this repo, https://github.com/AlexeyAB/darknet/blob/master/src/image.c, but I guess it's different. Please anybody has solved this?

[xinyee1997]

Can anyone help?

[larrywal-express]

can someone provide me with an edited image.c that converted rectangle box to circle box detection..

[Sceptersax]

Hello. This code is exactly what I need.. but after reading the entire thread it seems some sort of update has knocked the code out of commission? Any luck? anybody?

[MuhammadAsadJaved]

The code is changed recently I’ll try to send you old version.

On Sun, Jun 7, 2020 at 5:37 AM Sceptersax [e-mail] wrote:

Hello. This code is exactly what I need.. but after reading the entire
thread it seems some sort of update has knocked the code out of commission?
Any luck? anybody?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/AlexeyAB/darknet/issues/954#issuecomment-640121311,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/AG4GR5CLZQSY46YI5KJGZ3TRVKZIJANCNFSM4FCVU2EA
.

[MuhammadAsadJaved]

Give me your email address please

On Sun, Jun 7, 2020 at 5:37 AM Sceptersax [e-mail] wrote:

Hello. This code is exactly what I need.. but after reading the entire
thread it seems some sort of update has knocked the code out of commission?
Any luck? anybody?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/AlexeyAB/darknet/issues/954#issuecomment-640121311,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/AG4GR5CLZQSY46YI5KJGZ3TRVKZIJANCNFSM4FCVU2EA
.

[larrywal-express]

Here is my email for anyone that have the converted rectangle box to circle box code for image.c. thanks. [email protected]

[MuhammadAsadJaved]

It should not be that difficult, just change the “draw rectangular “
function to draw circle accordingly. You can see opencv draw circle
function.

On Sun, Jun 7, 2020 at 2:32 PM larrywal-express [e-mail]
wrote:

Here is my email for anyone that have the converted rectangle box to
circle box code for image.c. thanks. [email protected]

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/AlexeyAB/darknet/issues/954#issuecomment-640165209,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/AG4GR5CXJYDOB433HPKQBP3RVMX6NANCNFSM4FCVU2EA
.

[larrywal-express]

I mean for the draknet folder for Yolo. the image.c file inside the src

[MuhammadAsadJaved]

I also said same. You just need to change “cv2rectangular function to
cv2circle . Keep other things same.

You can also try this project, it’s relatively easy to change

https://github.com/ultralytics/yolov3

On Sun, Jun 7, 2020 at 2:54 PM larrywal-express [e-mail]
wrote:

I mean for the draknet folder for Yolo. the image.c file inside the src

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/AlexeyAB/darknet/issues/954#issuecomment-640167398,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/AG4GR5HPW2ILGIVKYYVF2NDRVM2RFANCNFSM4FCVU2EA
.

[larrywal-express]

Thanks. can you provide me the location of where to change? because I wouldn't see the cv2.rectangle function in image.c or you you can provide me where to change in the recommend project.

[Sceptersax]

@ MuhammadAsadJaved Hello,

I could still use the code. I attempted to contact you but didn't hear back. My email is:

[email protected]

Thanks

[MuhammadAsadJaved]

@ MuhammadAsadJaved Hello,

I could still use the code. I attempted to contact you but didn't hear back. My email is:

[email protected]

Thanks

OK. I will send you image.c file, you can try to use this image.c file.

Note: Do not delete original image.c file, copy or rename it for backup.
I tested it only on image input, not on video.

[larrywal-express]

thanks. I send it to your email.

[Sceptersax]

Will do. Thanks

On Sun, Jun 14, 2020 at 10:54 PM Asad [e-mail] wrote:

@ MuhammadAsadJaved Hello,

I could still use the code. I attempted to contact you but didn't hear
back. My email is:

[email protected]

Thanks

OK. I will send you image.c file, you can try to use this image.c file.

Note: Do not delete original image.c file, copy or rename it for backup.
I tested it only on image input, not on video.

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
https://github.com/AlexeyAB/darknet/issues/954#issuecomment-643918204,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/AC2UHTBVKH3UBYZ52VDWN2TRWWZPZANCNFSM4FCVU2EA
.

[Sceptersax]

I had some issues with my setup and couldn’t recompile the make file so I had to reinstall everything.. now I’m having other issues..

I wish to implement the cropping implementation into the image.c file but now I’m thinking I have the wrong opencv version. Currently I have opencv 3.6.

What version should I have?

[neheteakshay]

@AlexeyAB i have cropped detetcted part successfully and rsult is stored in result_img but when i am trying to show that image using opencv its giving me error as : AttributeError: 'NoneType' object has no attribute 'clip'
image exist but showing me this error. image size is small just 2.4kb only and dim 89x76.
what should i do?

[harshagarg09]

In order to crop multiple objects having different labels in a single image, replace "image.c" file in "src" folder with my image.c file. It will crop the detected objects and save them in .jpg format corresponding to different labels
image.zip

[Tanmaypal]

@halimB8 Hi,

1. Just un-comment these lines: https://github.com/AlexeyAB/darknet/blob/a720bf74e233c956e5281871ede722a2c7cf54db/src/image.c#L520-L534
2. Recompile Darknet
3. Create result_img directory near with darknet-executable file
4. Run darknet on the video file:
   ./darknet detector demo data/coco.data cfg/yolov3.cfg yolov3.weights test.mp4
5. Look at the result_img directory
   You will see files like img_<frame_id>_<img_id>_<class_id>.jpg with cropped detected objects.

About it: #846 (comment)

i am still not getting images in the result_img folder..plz help me out

[Tanmaypal]

sorry for the earlier query..
I am still not getting the images in the result_img folder..plz help me out

[headnut]

Works great for saving the jpg images in result_img, though png is desired.
For now added a line to print the bounding box coordinates, to allow for python post-processing.
Would be much easier for darknet to output them.

That was easy. Changed from
save_image(cropped_im, image_name);
to
save_image_png(cropped_im, image_name);